<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Future of Life on Medium]]></title>
        <description><![CDATA[Stories by Future of Life on Medium]]></description>
        <link>https://medium.com/@FLIxrisk?source=rss-e33e2d2a809c------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*2ZHeGJ4y2js6Rl2J.jpeg</url>
            <title>Stories by Future of Life on Medium</title>
            <link>https://medium.com/@FLIxrisk?source=rss-e33e2d2a809c------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Tue, 17 Nov 2026 17:34:17 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@FLIxrisk/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[FLI Podcast: AI Breakthroughs and Challenges in 2018 with David Krueger and Roman Yampolskiy]]></title>
            <link>https://medium.com/@FLIxrisk/fli-podcast-ai-breakthroughs-and-challenges-in-2018-with-david-krueger-and-roman-yampolskiy-b49fcdb0f20a?source=rss-e33e2d2a809c------2</link>
            <guid isPermaLink="false">https://medium.com/p/b49fcdb0f20a</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Future of Life]]></dc:creator>
            <pubDate>Thu, 31 Jan 2019 17:48:23 GMT</pubDate>
            <atom:updated>2019-02-05T21:13:04.135Z</atom:updated>
            <content:encoded><![CDATA[<iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fw.soundcloud.com%2Fplayer%2F%3Furl%3Dhttps%253A%252F%252Fapi.soundcloud.com%252Ftracks%252F567990117%26show_artwork%3Dtrue&amp;url=https%3A%2F%2Fsoundcloud.com%2Ffutureoflife%2Fai-breakthroughs-and-challenges-in-2018-with-david-krueger-and-roman-yampolskiy&amp;image=http%3A%2F%2Fi1.sndcdn.com%2Fartworks-000481289454-0txmm7-t500x500.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=soundcloud" width="800" height="166" frameborder="0" scrolling="no"><a href="https://medium.com/media/f126b2303493c51e873096e13f40831a/href">https://medium.com/media/f126b2303493c51e873096e13f40831a/href</a></iframe><p>Every January, we like to look back over the past 12 months at the progress that’s been made in the world of artificial intelligence. Welcome to our annual “AI breakthroughs” podcast, 2018 edition.</p><p>Ariel was joined for this retrospective by researchers Roman Yampolskiy and David Krueger. Roman is an AI Safety researcher and professor at the University of Louisville. He also recently published the book <a href="https://www.amazon.com/Artificial-Intelligence-Security-Chapman-Robotics/dp/0815369824"><em>Artificial Intelligence Safety &amp; Security</em></a>. David is a PhD candidate in the Mila lab at the University of Montreal, where he works on deep learning and AI safety. He’s also worked with safety teams at the Future of Humanity Institute and DeepMind and has volunteered with 80,000 hours.</p><p>Roman and David shared their lists of 2018’s most promising AI advances, as well as their thoughts on some major ethical questions and safety concerns. They also discussed media coverage of AI research, why talking about “breakthroughs” can be misleading, and why there may have been more progress in the past year than it seems.</p><p>Topics discussed in this podcast include:</p><ul><li>DeepMind progress, as seen with <a href="https://deepmind.com/blog/alphastar-mastering-real-time-strategy-game-starcraft-ii/">AlphaStar</a> and <a href="https://deepmind.com/blog/alphafold/">AlphaFold</a></li><li>Manual dexterity in robots, especially <a href="https://arxiv.org/abs/1806.10293">QT Opt</a> and <a href="https://www.theverge.com/2018/7/30/17621112/openai-robot-dexterity-dactyl-artificial-intelligence">Dactyl</a></li><li>Advances in creativity, as with Generative Adversarial Networks (GANs)</li><li><a href="https://distill.pub/2018/feature-wise-transformations/">Feature-wise transformations</a></li><li>Continuing concerns about DeepFakes</li><li>Scaling up AI systems</li><li>Neuroevolution</li><li><a href="https://ai.googleblog.com/2018/05/duplex-ai-system-for-natural-conversation.html">Google Duplex</a>, the AI assistant that sounds human on the phone</li><li>The General Data Protection Regulation (GDPR) and AI policy more broadly</li></ul><p>Publications discussed in this podcast include:</p><ul><li><a href="https://arxiv.org/pdf/1810.08575.pdf">Supervising strong learners by amplifying weak experts</a> (OpenAI)</li><li><a href="https://arxiv.org/pdf/1805.00899.pdf">AI safety via debate</a> (OpenAI)</li><li><a href="https://arxiv.org/pdf/1811.07871.pdf">Scalable agent alignment via reward modeling: a research direction</a> (DeepMind)</li><li><a href="https://blog.openai.com/ai-and-compute/">AI and Compute</a> (OpenAI)</li><li><a href="https://arxiv.org/abs/1812.04948">A Style-Based Generator Architecture for Generative Adversarial Networks</a></li><li><a href="https://arxiv.org/abs/1802.01561">IMPALA: Scalable Distributed Deep-RL with Importance Weighted Actor-Learner Architectures</a></li><li><a href="https://arxiv.org/abs/1811.05233">ImageNet/ResNet-50 Training in 224 Seconds</a></li><li><a href="https://arxiv.org/abs/1706.03762">Attention Is All You Need</a></li></ul><p>You can listen to the podcast above, or read the full transcript below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*dk3scK1WcYUOsgiLrNPvYw.jpeg" /></figure><p><strong>Ariel:</strong> Hi everyone, welcome to the FLI podcast. I’m your host, Ariel Conn. For those of you who are new to the podcast, at the end of each month, I bring together two experts for an in-depth discussion on some topic related to the fields that we at the Future of Life Institute are concerned about, namely artificial intelligence, biotechnology, climate change, and nuclear weapons.</p><p>The last couple of years for our January podcast, I’ve brought on two AI researchers to talk about what the biggest AI breakthroughs were in the previous year, and this January is no different. To discuss the major developments we saw in AI in 2018, I’m pleased to have Roman Yampolskiy and David Krueger joining us today.</p><p>Roman is an AI safety researcher and professor at the University of Louisville, his new book <em>Artificial Intelligence Safety and Security</em> is now available on Amazon and we’ll have links to it on the FLI page for this podcast. David is a PhD candidate in the Mila Lab at the University of Montreal, where he works on deep learning and AI safety. He’s also worked with teams at the Future of Humanity Institute and DeepMind, and he’s volunteered with 80,000 Hours to help people find ways to contribute to the reduction of existential risks from AI. So Roman and David, thank you so much for joining us.</p><p><strong>David:</strong> Yeah, thanks for having me.</p><p><strong>Roman:</strong> Thanks very much.</p><p><strong>Ariel:</strong> So I think that one thing that stood out to me in 2018 was that the AI breakthroughs seemed less about surprising breakthroughs that really shook the AI community as we’ve seen in the last few years, and instead they were more about continuing progress. And we also didn’t see quite as many major breakthroughs hitting the mainstream press. There were a couple of things that made big news splashes, like Google Duplex, which is a new AI assistant program that sounded incredibly human on phone calls it made during the demos. And there was also an uptick in government policy and ethics efforts, especially with the General Data Protection Regulation, also known as the GDPR, which went into effect in Europe earlier this year.</p><p>Now I’m going to want to come back to Google and policy and ethics later in this podcast, but I want to start by looking at this from the research and development side of things. So my very first question for both of you is: do you agree that 2018 was more about impressive progress, and less about major breakthroughs? Or were there breakthroughs that really were important to the AI community that just didn’t make it into the mainstream press?</p><p><strong>David:</strong> Broadly speaking I think I agree, although I have a few caveats for that. One is just that it’s a little bit hard to recognize always what is a breakthrough, and a lot of the things in the past that have had really big impacts didn’t really seem like some amazing new paradigm shift — it was sort of a small tweak that then made a lot of things work a lot better. And the other caveat is that there are a few works that I think are pretty interesting and worth mentioning, and the field is so large at this point that it’s a little bit hard to know if there aren’t things that are being overlooked.</p><p><strong>Roman:</strong> So I’ll agree with you, but I think the pattern is more important than any specific breakthrough. We kind of got used to getting something really impressive every month, so relatively it doesn’t sound as good, all the AlphaStar, AlphaFold, AlphaZero happening almost every month. And it used to be it took 10 years to see something like that.</p><p>It’s likely it will happen even more frequently. We’ll conquer a new domain once a week or something. I think that’s the main pattern we have to recognize and discuss. There are significant accomplishments in terms of teaching AI to work in completely novel domains. I mean now we can predict protein folding, now we can have multi-player games conquered. That never happened before so frequently. Chess was impressive because it took like 30 years to get there.</p><p><strong>David:</strong> Yeah, so I think a lot of people were kind of expecting or at least hoping for StarCraft or Dota to be solved — to see, like we did with AlphaGo, AI systems that are beating the top players. And I would say that it’s actually been a little bit of a let down for people who are optimistic about that, because so far the progress has been kind of unconvincing.</p><p>So the AlphaStar, which was a really recent result from last week, for instance: I’ve seen criticism of it that I think is valid that it was making more actions than a human could within a very short interval of time. So they carefully controlled the actions-per-minute that AlphaStar was allowed to take, but they didn’t prevent it from doing really short bursts of actions that really helped its micro-game, and that means that it can win without really being strategically superior to its human opponents. And I think the Dota results that OpenAI has had was also criticized as being sort of not the hardest version of the problem, and still the AI sort of is relying on some crutches.</p><p><strong>Ariel:</strong> So before we get too far into that debate, can we take a quick step back and explain what both of those are?</p><p><strong>David:</strong> So these are both real-time strategy games that are, I think, actually the two most popular real-time strategy games in the world that people play professionally, and make money playing. I guess that’s all to say about them.</p><p><strong>Ariel:</strong> So a quick question that I had too about your description then, when you’re talking about AlphaStar and you were saying it was just making more moves than a person can realistically make. Is that it — it wasn’t doing anything else special?</p><p><strong>David:</strong> I haven’t watched the games, and I don’t play StarCraft, so I can’t say that it wasn’t doing anything special. I’m basing this basically on reading articles and reading the opinions of people who are avid StarCraft players, and I think the general opinion seems to be that it is more sophisticated than what we’ve seen before, but the reason that it was able to win these games was not because it was out-thinking humans, it’s because it was out-clicking, basically, in a way that just isn’t humanly possible.</p><p><strong>Roman:</strong> I would agree with this analysis, but I don’t see it as a bug, I see it as a feature. That just shows another way machines can be superior to people. Even if they are not necessarily smarter, they can still produce superior performance, and that’s what we really care about. Right? We found a different way, a non-human approach to solving this problem. That’s impressive.</p><p><strong>David:</strong> Well, I mean, I think if you have an agent that can just click as fast as it wants, then you can already win at StarCraft, before this work. There needs to be something that makes it sort of a fair fight in some sense.</p><p><strong>Roman:</strong> Right, but think what you’re suggesting: We have to handicap machines to make them even remotely within being comparative to people. We’re talking about getting to superintelligent performance. You can get there by many ways. You can think faster, you can have better memory, you can have better reaction time — as long as you’re winning in whatever domain we’re interested in, you have superhuman performance.</p><p><strong>David Krueger:</strong> So maybe another way of putting this would be if they actually made a robot play StarCraft and made it use the same interface that humans do, such as a screen and mouse, there’s no way that it could have beat the human players. And so by giving it direct access to the game controls, it’s sort of not solving the same problem that a human is when they play this game.</p><p><strong>Roman:</strong> I feel what you’re saying, I just feel that it is solving it in a different way, and we have pro-human bias saying, well that’s not how you play this game, you have an advantage. Human players usually rely on superior strategy, not just faster movements that may take advantage of it for a few nanoseconds, a couple of seconds. But it’s not a long-term sustainable pattern.</p><p>One of the research projects I worked on was this idea of artificial stupidity, we called it — kind of limiting machines to human-level capacity. And I think that’s what we’re talking about it here. Nobody would suggest limiting a chess program to just human-level memory, or human memorization of opening moves. But we don’t see it as a limitation. Machines have an option of beating us in ways humans can’t. That’s the whole point, and that’s why it’s interesting, that’s why we have to anticipate such problems. That’s where most of the safety and security issues will show up.</p><p><strong>Ariel:</strong> So I guess, I think, Roman, your point earlier was sort of interesting that we’ve gotten so used to breakthroughs that stuff that maybe a couple of years ago would have seemed like a huge breakthrough is just run-of-the-mill progress. I guess you’re saying that that’s what this is sort of falling into. Relatively recently this would have been a huge deal, but because we’ve seen so much other progress and breakthroughs, that this is now interesting and we’re excited about it — but it’s not reaching that level of, oh my god, this is amazing! Is that fair to say?</p><p><strong>Roman:</strong> Exactly! We get disappointed if the system loses one game. It used to be we were excited if it would match amateur players. Now it’s, oh, we played a 100 games and you lost one? This is just not machine-level performance, you disappoint us.</p><p><strong>Ariel:</strong> David, do you agree with that assessment?</p><p><strong>David:</strong> I would say mostly no. I guess, I think what really impressed me with AlphaGo and AlphaZero was that it was solving something that had been established as a really grand challenge for AI. And then in the case of AlphaZero, I think the technique that they actually used to solve it was really novel and interesting from a research point of view, and they went on to show that this same technique can solve a bunch of other board games as well.</p><p>And my impression from what I’ve seen about how they did AlphaStar and AlphaFold is that there were some interesting improvements and the performance is impressive but I think it’s neither, like, quite at the point where you can say we’ve solved it, we’re better than everybody, or in the case of protein folding, there’s not a bunch more room for improvement that has practical significance. And it’s also — I don’t see any really clear general algorithmic insights about AI coming out of these works yet. I think that’s partially because they haven’t been published yet, but from what I have heard about the details about how they work, I think it’s less of a breakthrough on the algorithm side than AlphaZero was.</p><p><strong>Ariel:</strong> So you’ve mentioned AlphaFold. Can you explain what that is real quick?</p><p><strong>David:</strong> This is the protein folding project that DeepMind did, and I think there’s a competition called C-A-S-P or CASP that happens every three years, and they sort of dominated that competition this last year doing what was described as two CASPs in one, so basically doubling the expected rate of improvement that people have seen historically at these tasks, or at least at the one that is the most significant benchmark.</p><p><strong>Ariel:</strong> I find the idea of the protein folding thing interesting because that’s something that’s actually relevant to scientific advancement and health as opposed to just being able to play a game. Are we seeing actual applications for this yet?</p><p><strong>David:</strong> I don’t know about that, but I agree with you that that is a huge difference that makes it a lot more exciting than some of the previous examples. I guess one thing that I want to say about that, though, is that it does look a little bit more to me like continuation of progress that was already happening in the communities. It’s definitely a big step up, but I think a lot of the things that they did there could have really happened over the next few years anyways, even without DeepMind being there. So, one of the articles I read put it this way: If this wasn’t done by DeepMind, if this was just some academic group, would this have been reported in the media? I think the answer is sort of like a clear no, and that says something about the priorities of our reporting and media as well as the significance of the results, but I think that just gives some context.</p><p><strong>Roman:</strong> I’ll agree with David — the media is terrible in terms of what they report on, we can all agree on that. I think it was quite a breakthrough, I mean, to say that they not just beat the competition, but to actually kind of doubled performance improvement. That’s incredible. And I think anyone who got to that point would not be denied publication in a top journal; It would be considered very important in that domain. I think it’s one of the most important problems in medical research. If you can accurately predict this, possibilities are really endless in terms of synthetic biology, in terms of curing diseases.</p><p>So this is huge in terms of impact from being able to do it. As far as how applicable is it to other areas, is it a great game-changer for AI research? All those things can adapt between this ability to perform in real-life environments of those multiplayer games, and being able to do this. Look at how those things can be combined. Right? You can do things in the real world you couldn’t do before, both in terms of strategy games, which are basically simulations for economic competition, for wars, for quite a few applications where impact would be huge.</p><p>So all of it is very interesting. It’s easy to say that, “Well if they didn’t do it, somebody else maybe would do it in a couple of years.” But it’s almost always true for all inventions. If you look at the history of inventions, things like, I don’t know, telephone, have been invented at the same time by two or three people; radio, two or three people. It’s just the point where science gets enough ingredient technology where yeah, somebody’s going to do it, nice. But still, we give credit to whoever got there first.</p><p><strong>Ariel:</strong> So I think that’s actually a really interesting point, because I think for the last few years we have seen sort of these technological advances but I guess we also want to be considering the advances that are going to have a major impact on humanity even if it’s not quite as technologically new.</p><p><strong>David:</strong> Yeah, absolutely. I think the framing in terms of breakthroughs is a little bit unclear what we’re talking about when we talk about AI breakthroughs, and I think a lot of people in the field of AI kind of don’t like how much people talk about it in terms of breakthroughs because a lot of the progress is gradual and builds on previous work and it’s not like there was some sudden insight that somebody had that just changed everything, although that does happen in some ways.</p><p>And I think you can think of the breakthroughs both in terms of like what is the impact — is this suddenly going to have a lot of potential to change the world? You can also think of it, though, from the perspective of researchers as like, is this really different from the kind of ideas and techniques we’ve seen or seen working before? I guess I’m more thinking about the second right now in terms of breakthroughs representing really radical new ideas in research.</p><p><strong>Ariel:</strong> Okay, well I will take responsibility for being one of the media people who didn’t do a good job with presenting AI breakthroughs. But I think both with this podcast and probably moving forward, I think that is actually a really important thing for us to be doing — is both looking at the technological progress and newness of something but also the impact it could have on either society or future research.</p><p>So with that in mind, you guys also have a good list of other things that did happen this year, so I want to start moving into some of that as well. So next on your list is manual dexterity in robots. What did you guys see happening there?</p><p><strong>David:</strong> So this is something that’s definitely not my area of expertise, so I can’t really comment too much on it. But there are two papers that I think are significant and potentially representing something like a breakthrough in this application. In general robotics is really difficult, and machine learning for robotics is still, I think, sort of a niche thing, like most robotics is using more classical planning algorithms, and hasn’t really taken advantage of the new wave of deep learning and everything.</p><p>So there’s two works, one is QT-Opt, and the other one is Dactyl, and these are both by people from the Berkeley OpenAI crowd. And these both are showing kind of impressive results in terms of manual dexterity in robots. So there’s one that does a really good job at grasping, which is one of the basic aspects of being able to act in the real world. And then there’s another one that was sort of just manipulating something like a cube with different colored faces on it — that one’s Dactyl; the grasping one is QT-Opt.</p><p>And I think this is something that was paid less attention to in the media, because it’s been more of a story of kind of gradual progress I think. But my friend who follows this deep reinforcement learning stuff more told me that QT-Opt is the first convincing demonstration of deep reinforcement learning in the real world, as opposed to all these things we’ve seen in games. The real world is much more complicated and there’s all sorts of challenges with the noise of the environment dynamics and contact forces and stuff like this that have been really a challenge for doing things in the real world. And then there’s also the limited sample complexity where when you play a game you can sort of interact with the game as much as you want and play the game over and over again, whereas in the real world you can only move your robot so fast and you have to worry about breaking it, so that means in the end you can collect a lot less data, which makes it harder to learn things.</p><p><strong>Roman:</strong> Just to kind of explain maybe what they did. So hardware’s expensive, slow: It’s very difficult to work with. Things don’t go well in real life; It’s a lot easier to create simulations in virtual worlds, train your robot in there, and then just transfer knowledge into a real robot in the physical world. And that’s exactly what they did, training that virtual hand to manipulate objects, and they could run through thousands, millions of situations and then it’s something you cannot do with an actual, physical robot at that scale. So, I think that’s a very interesting approach for why lots of people try doing things in virtual environments. Some of the early AGI projects all concentrated on virtual worlds as domain of learning. So that makes a lot of sense.</p><p><strong>David:</strong> Yeah, so this was for the Dactyl project, which was OpenAI. And that was really impressive I think, because people have been doing this sim-to-real thing — where you train in simulation and then try and transfer it to the real world — with some success for like a year or two, but this one I think was really kind of impressive in that sense, because they didn’t actually train it in the real world at all, and what they had learned managed to transfer to the real world.</p><p><strong>Ariel:</strong> Excellent. I’m going to keep going through your list. One thing that you both mentioned are GANs. So very quickly, if one of you, or both of you, could explain what a GAN is and what that stands for, and then we’ll get into what happened last year with those.</p><p><strong>Roman:</strong> Sure, so this is a somewhat new way of doing creative generational visuals and audio. You have two neural networks competing, one is kind of creating fakes, and the other one is judging them, and you get to a point where they’re kind of 50/50. You can’t tell if it’s fake or real anymore. And it’s a great way to produce artificial faces, cars, whatever. Any type of input you can provide to the networks, they quickly learn to extract the essence of that image or audio and generate artificial data sets full of such images.</p><p>And there’s really exciting work on being able to extract properties from those, different styles. So if we talk about faces, for example: there could be a style for hair, a style for skin color, a style for age, and now it’s possible to manipulate them. So I can tell you things like, “Okay, Photoshop, I need a picture of a female, 20 years old, blonde, with glasses,” and it would generate a completely realistic face based on those properties. And we’re starting to see it show up not just in images but transferred to video, to generating whole virtual worlds. It’s probably the closest thing we ever had computers get to creativity: actually kind of daydreaming and coming up with novel outputs.</p><p><strong>David:</strong> Yeah, I just want to say a little bit about the history of the research in GAN. So the first work on GANs was actually back four or five years ago in 2014, and I think it was actually kind of — didn’t make a huge splash at the time, but maybe a year or two after that it really started to take off. And research in GANs over the last few years has just been incredibly fast-paced and there’s been hundreds of papers submitted and published at the big conferences every year.</p><p>If you look just in terms of the quality of what is generated, this is, I think, just an amazing demonstration of the rate of progress in some areas of machine learning. The first paper had these sort of black and white pictures of really blurry faces, and now you can get giant — I think 256 by 256, or 512 by 512, or even bigger — really high resolution and totally indistinguishable from real photos, to the human eye anyway — images of faces. So it’s really impressive, and we’ve seen really consistent progress on that, especially in the last couple years.</p><p><strong>Ariel:</strong> And also, just real quick, what does it stand for?</p><p><strong>David:</strong> Oh, generative adversarial network. So it’s generative, because it’s sort of generating things from scratch, or from its imagination or creativity. And it’s adversarial because there are two networks: the one that generates the things, and then the one that tries to tell those fake images apart from real images that we actually collect by taking photos in the world.</p><p><strong>Ariel:</strong> This is an interesting one because it can sort of transition into some ethics stuff that came up this past year, but I’m not sure if we want to get there yet, or if you guys want to talk a little bit more about some of the other things that happened on the research and development side.</p><p><strong>David:</strong> I guess I want to talk about a few other things that have been making, I would say, sort of steady progress, like GANs. With a lot of interest in, I guess I would say, their ideas that are coming to fruition, even though some of these are not exactly from the last year, they sort of really started to improve themselves and become widely used in the last year.</p><p><strong>Ariel:</strong> Okay.</p><p><strong>David:</strong> I think this is actually used in maybe the latest, greatest GAN paper, is something that’s called feature-wise transformations. So this is an idea that actually goes back up to 40 years, depending on how you measure it, but has sort of been catching on in specific applications in machine learning in the last couple of years — starting with, I would say, style-transfer, which is sort of like what Roman mentioned earlier.</p><p>So the idea here is that in a neural network, you have what are called features, which basically correspond to the activations of different neurons in the network. Like how much that neuron likes what it’s seeing, let’s say. And those can also be interpreted as representing different kinds of visual patterns, like different kinds of textures, or colors. And these feature-wise transformations basically just take each of those different aspects of the image, like the color or texture in a certain location, and then allow you to manipulate that specific feature, as we call it, by making it stronger or amplifying whatever was already there.</p><p>And so you can sort of view this as a way of specifying what sort of things are important in the image, and that’s why it allows you to manipulate the style of images very easily, because you can sort of look at a certain painting style for instance, and say, oh this person uses a lot of wide brush strokes, or a lot of narrow brush strokes, and then you can say, I’m just going to modulate the neurons that correspond to wide or narrow brush strokes, and change the style of the painting that way. And of course you don’t do this by hand, by looking in and seeing what the different neurons represent. This all ends up being learned end-to-end. And so you sort of have an artificial intelligence model that predicts how to modulate the features within another network, and that allows you to change what that network does in a really powerful way.</p><p>So, I mentioned that it has been applied in the most recent GAN papers, and I think they’re just using those kinds of transformations to help them generate images. But other examples where you can explain what’s happening more intuitively, or why it makes sense to try and do this, would be something like visual question answering. So there you can have the modulation of the vision network being done by another network that looks at a question and is trying to help answer that question. And so it can sort of read the question and see what features of images might be relevant to answering that question. So for instance, if the question was, “Is it a sunny day outside?” then it could have the vision network try and pay more attention to things that correspond to signs of sun. Or if it was asked something like, “Is this person’s hair combed?” then you could look for the patterns of smooth, combed hair and look for the patterns of rough, tangled hair, and have those features be sort of emphasized in the vision network. That allows the vision network to pay attention to the parts of the image that are most relevant to answering the question.</p><p><strong>Ariel:</strong> Okay. So, Roman, I want to go back to something on your list quickly in a moment, but first I was wondering if you have anything that you wanted to add to the feature-wise transformations?</p><p><strong>Roman:</strong> All of it, you can ask, “Well why is this interesting, what are the applications for it?” So you are able to generate inputs, inputs for computers, inputs for people, images, sounds, videos. A lot of times they can be adversarial in nature as well — what we call deep fakes. Right? You can make, let’s say, a video of a famous politician say something, or do something.</p><p><strong>Ariel:</strong> Yeah.</p><p><strong>Roman:</strong> And this has very interesting implications for elections, for forensic science, for evidence. As those systems get better and better, it becomes harder and harder to tell if something is real or not. And maybe it’s still possible to do some statistical analysis, but it takes time, and we talked about media being not exactly always on top of it. So it may take 24 hours before we realize if this video was real or not, but the election is tonight.</p><p><strong>Ariel:</strong> So I am definitely coming back to that. I want to finish going through the list of the technology stuff, but yeah I want to talk about deep fakes and in general, a lot of the issues that we’ve seen cropping up more and more with this idea of using AI to fake images and audio and video, because I think that is something that’s really important.</p><p><strong>David:</strong> Yeah, it’s hard for me to estimate these things, but I would say this is probably, in terms of the impact that this is going to have societally, this is sort of the biggest story maybe of the last year. And it’s not like something that happened all of the sudden. Again, it’s something that has been building on a lot of progress in generative models and GANs and things like this. And it’s just going to continue, we’re going to see more and more progress like that, and probably some sort of arms’ race here where — I shouldn’t use that word.</p><p><strong>Ariel:</strong> A competition.</p><p><strong>David:</strong> A competition between people who are trying to use that kind of technology to fake things and people who are sort of doing forensics to try and figure out what is real and what is fake. And that also means that people are going to have to trust the people who have the expertise to do that, and believe that they’re actually doing that and not part of some sort of conspiracy or something.</p><p><strong>Ariel:</strong> Alright, well are you guys ready to jump into some of those ethical questions?</p><p><strong>David:</strong> Well, there are like two other broad things I wanted to mention, which I think are sort of interesting trends in the research community. One is just the way that people have been continuing to scale up AI systems. So a lot of the progress I think has arguably just been coming from more and more computation and more and more data. And there was a pretty great blog post by OpenAI about this last year that argued that the amount of computation that’s being used to train the most advanced AI systems is increasing by a factor of 10 times every year for the last several years, which is just astounding. But it also suggests that this might not be sustainable for a long time, so to the extent that you think that using more computation is a big driver of progress, we might start to see that slow down within a decade or so.</p><p><strong>Roman:</strong> I’ll add another — what I think also is kind of building-on technology, not so much a breakthrough, we had it for a long time — but neural evolution is something I’m starting to pay a lot more attention to and that’s kind of borrowing from biology, trying to evolve ways for neural networks, optimized neural networks. And it’s producing very impressive results. It’s possible to run it in parallel really well, and it’s competitive with some of the leading alternative approaches.</p><p>So, the idea basically is you have this very large neural network, brain-like structure, but instead of trying to train it back, propagate errors, teach it in a standard neural networks way, you just kind of have a population of those brains competing for who’s doing best in a particular problem, and they share weights between good parents, and after a while you just evolve really well performing solutions to some of the most interesting problems.</p><p>Additionally you can kind of go meta-level on it and evolve architectures for the neural network itself — how many layers, how many inputs. This is nice because it doesn’t require much human intervention. You’re essentially letting the system figure out what the solutions are. We had some very successful results with genetic algorithms for optimization. We didn’t have much success with genetic programming, and now neural evolution kind of brings it back where you’re optimizing intelligence systems, and that’s very exciting.</p><p><strong>Ariel:</strong> So you’re saying that you’ll have — to make sure I understand this correctly — there’s two or more neural nets trying to solve a problem, and they sort of play off of each other?</p><p><strong>Roman:</strong> So you create a population of neural networks, and you give it a problem, and you see this one is doing really well, and that one. The others, maybe not so great. So you take weights from those two and combine them — like mom and dad, parent situation that produces offspring. And so you have this simulation of evolution where unsuccessful individuals are taken out of a population. Successful ones get to reproduce and procreate, and provide their high fitness weights to the next generation.</p><p><strong>Ariel:</strong> Okay. Was there anything else that you guys saw this year that you want to talk about, that you were excited about?</p><p><strong>David:</strong> Well I wanted to give a few examples of the kind of massive improvements in scale that we’ve seen. One of the most significant models and benchmarks in the community is ImageNet and training image classifiers that can tell you what a picture is a picture of on this dataset.So the whole sort of deep learning revolution was arguably started, or at least really came into the eyes of the rest of the machine learning community, because of huge success on this ImageNet competition. And training the model there took something like two weeks, and this last year there was a paper where you can train a more powerful model in less than four minutes, and they do this by using like 3000 graphics cards in parallel.</p><p>And then DeepMind also had some progress on parallelism with this model called IMPALA, which basically was in the context of reinforcement learning as opposed to classification, and there they sort of came up with a way that allowed them to do updates in parallels, like learn on different machines and combine everything that was learned in a way that’s asynchronous. So in the past the sort of methods that they would use for these reinforcement learning problems, you’d have to wait for all of the different machines to finish their learning on the current problem or instance that they’re learning about, and then combine all of that centrally — whereas the new method allows you to just as soon as you’re done computing or learning something, you can communicate it to the rest of the system, the other computers that are learning in parallel. And that was really important for allowing them to scale to hundreds of machines working on their problem at the same time.</p><p><strong>Ariel:</strong> Okay, and so that, just to clarify as well, that goes back to this idea that right now we’re seeing a lot of success just scaling up the computing, but at some point that could slow things down essentially, if we had a limit for how much computing is possible.</p><p><strong>David:</strong> Yeah, and I guess one of my points is also doing this kinds of scaling of computing requires some amount of algorithmic insight or breakthrough if you want to be dramatic as well. So this DeepMind paper I talked about, they had to devise new reinforcement learning algorithms that would still be stable when they had this real-time asynchronous updating. And so, in a way, yeah, a lot of the research that’s interesting right now is on finding ways to make the algorithm scale so that you can keep taking advantage of more and more hardware. And the evolution stuff also fits into that picture to some extent.</p><p><strong>Ariel:</strong> Okay. I want to start making that transition into some of the concerns that we have for misuse around AI and how easy it is for people to be deceived by things that have been created by AI. But I want to start with something that’s hopefully a little bit more neutral, and talk about Google Duplex, which is the program that Google came out with, I think last May. I don’t know the extent to which it’s in use now, but they presented it, and it’s an AI assistant that can essentially make calls and set up appointments for you. So their examples were it could make a reservation at a restaurant for you, or it could make a reservation for you to get a haircut somewhere. And it got sort of mixed reviews, because on the one hand people were really excited about this, and on the other hand it was kind of creepy because it sounded human, and the people on the other end of the call did not know that they were talking to a machine.</p><p>So I was hoping you guys could talk a little bit I guess maybe about the extent to which that was an actual technological breakthrough versus just something — this one being more one of those breakthroughs that will impact society more directly. And then also I guess if you agree that this seems like a good place to transition into some of the safety issues.</p><p><strong>David:</strong> Yeah, no, I would be surprised if they really told us about the details of how that worked. So it’s hard to know how much of an algorithmic breakthrough or algorithmic breakthroughs were involved. It’s very impressive, I think, just in terms of what it was able to do, and of course these demos that we saw were maybe selected for their impressiveness. But I was really, really impressed personally, just to see a system that’s able to do that.</p><p><strong>Roman:</strong> It’s probably built on a lot of existing technology, but it is more about impact than what you can do with this. And my background is cybersecurity, so I see it as a great tool for like automating spear-phishing attacks on a scale of millions. You’re getting a real human calling you, talking to you, with access to your online data; Pretty much everyone’s gonna agree and do whatever the system is asking of you, if it’s credit card numbers, or social security numbers. So, in many ways it’s going to be a game changer.</p><p><strong>Ariel:</strong> So I’m going to take that as a definite transition into safety issues. So, yeah, let’s start talking about, I guess, sort of human manipulation that’s happening here. First, the phrase “deep fake” shows up a lot. Can you explain what those are?</p><p><strong>David:</strong> So “deep fakes” is basically just: you can make a fake video of somebody doing something or saying something that they did not actually do or say. People have used this to create fake videos of politicians, they’ve used it to create porn using celebrities. That was one of the things that got it on the front page of the internet, basically. And Reddit actually shut down the subreddit where people were doing that. But, I mean, there’s all sorts of possibilities.</p><p><strong>Ariel:</strong> Okay, so I think the Reddit example was technically the very end of 2017. But all of this sort became more of an issue in 2018. So we’re seeing this increase in capability to both create images that seem real, create audio that seems real, create video that seems real, and to modify existing images and video and audio in ways that aren’t immediately obvious to a human. What did we see in terms of research to try to protect us from that, or catch that, or defend against that?</p><p><strong>Roman:</strong> So here’s an interesting observation, I guess. You can develop some sort of a forensic tool to analyze it, and give you a percentage likelihood that it’s real or that it’s fake. But does it really impact people? If you see it with your own eyes, are you going to believe your lying eyes, or some expert statistician on CNN?</p><p>So the problem is it will still have tremendous impact on most people. We’re not very successful at convincing people about multiple scientific facts. They simply go outside, or it’s cold right now, so global warming is false. I suspect we’ll see exactly that with, let’s say, fake videos of politicians, where a majority of people easily believe anything they hear once or see once versus any number of peer reviewed publications disproving it.</p><p><strong>David:</strong> I kind of agree. I mean, I think, when I try to think about how we would actually solve this kind of problem, I don’t think a technical solution that just allows somebody who has technical expertise to distinguish real from fake is going to be enough. We really need to figure out how to build a better trust infrastructure in our whole society which is kind of a massive project. I’m not even sure exactly where to begin with that.</p><p><strong>Roman:</strong> I guess the good news is it gives you plausible deniability. If a video of me comes out doing horrible things I can play it straight.</p><p><strong>Ariel:</strong> That’s good for someone. Alright, so, I mean, you guys are two researchers, I don’t know how into policy you are, but I don’t know if we saw as many strong policies being developed. We did see the implementation of the GDPR, and for people who aren’t familiar with the GDPR, it’s essentially European rules about what data companies can collect from your interactions online, and the ways in which you need to give approval for companies to collect your data, and there’s a lot more to it than that. One of the things that I found most interesting about the GDPR is that it’s entirely European based, but it had a very global impact because it’s so difficult for companies to apply something only in Europe and not in other countries. And so earlier this year when you were getting all of those emails about privacy policies, that was all triggered by the GDPR. That was something very specific that happened and it did make a lot of news, but in general I felt that we saw a lot of countries and a lot of national and international efforts for governments to start trying to understand how AI is going to be impacting their citizens, and then also trying to apply ethics and things like that.</p><p>I’m sort of curious, before we get too far into anything: just as researchers, what is your reaction to that?</p><p><strong>Roman:</strong> So I never got as much spam as I did that week when they released this new policy, so that kind of gives you a pretty good summary of what to expect. If you look at history, we have regulations against spam, for example. Computer viruses are illegal. So that’s a very expected result. It’s not gonna solve technical problems. Right?</p><p><strong>David:</strong> I guess I like that they’re paying attention and they’re trying to tackle these issues. I think the way GDPR was actually worded, it has been criticized a lot for being either much too broad or demanding, or vague. I’m not sure — there are some aspects of the details of that regulation that I’m not convinced about, or not super happy about. I guess overall it seems like people who are making these kinds of decisions, especially when we’re talking about cutting edge machine learning, it’s just really hard. I mean, even people in the fields don’t really know how you would begin to effectively regulate machine learning systems, and I think there’s a lot of disagreement about what a reasonable level of regulation would be or how regulations should work.</p><p>People are starting to have that sort of conversation in the research community a little bit more, and maybe we’ll have some better ideas about that in a few years. But I think right now it seems premature to me to even start trying to regulate machine learning in particular, because we just don’t really know where to begin. I think it’s obvious that we do need to think about how we control the use of the technology, because it’s just so powerful and has so much potential for harm and misuse and accidents and so on. But I think how you actually go about doing that is a really unclear and difficult problem.</p><p><strong>Ariel:</strong> So for me it’s sort of interesting, we’ve been debating a bit today about technological breakthroughs versus societal impacts, and whether 2018 actually had as many breakthroughs and all of that. But I would guess that all of us agree that AI is progressing a lot faster than government does.</p><p><strong>David:</strong> Yeah.</p><p><strong>Roman:</strong> That’s almost a tautology.</p><p><strong>Ariel:</strong> So I guess as researchers, what concerns do you have regarding that? Like do you worry about the speed at which AI is advancing?</p><p><strong>David:</strong> Yeah, I would say I definitely do. I mean, we were just talking about this issue with fakes and how that’s going to contribute to things like fake news and erosion of trust in media and authority and polarization of society. I mean, if AI wasn’t going so fast in that direction, then we wouldn’t have that problem. And I think the rate that it’s going, I don’t see us catching up — or I should say, I don’t see the government catching up on its own anytime soon — to actually control the use of AI technology, and do our best anyways to make sure that it’s used in a safe way, and a fair way, and so on.</p><p>I think in and of itself it’s maybe not bad that the technology is progressing fast. I mean, it’s really amazing; Scientifically there’s gonna be all sorts of amazing applications for it. But there’s going to be more and more problems as well, and I don’t think we’re really well equipped to solve them right now.</p><p><strong>Roman:</strong> I’ll agree with David, I’m very concerned at its relative rate of progress. AI development progresses a lot faster than anything we see in AI safety. AI safety is just trying to identify problem areas, propose some general directions, but we have very little to show in terms of solved problems.</p><p>If you look at our work in adversarial fields, maybe a little bit cryptography, the good guys have always been a step ahead of the bad guys, whereas here you barely have any good guys as a percentage. You have like less than 1% of researchers working directly on safety full-time. Same situation with funding. So it’s not a very optimistic picture at this point.</p><p><strong>David:</strong> I think it’s worth definitely distinguishing the kind of security risks that we’re talking about, in terms of fake news and stuff like that, from long-term AI safety, which is what I’m most interested in, and think is actually even more important, even though I think there’s going to be tons of important impacts we have to worry about already, and in the coming years.</p><p>And the long-term safety stuff is really more about artificial intelligence that becomes broadly capable and as smart or smarter than humans across the board. And there, there’s maybe a little bit more signs of hope if I look at how the fields might progress in the future, and that’s because there’s a lot of problems that are going to be relevant for controlling or aligning or understanding these kind of generally intelligent systems that are probably going to be necessary anyways in terms of making systems that are more capable in the near future.</p><p>So I think we’re starting to see issues with trying to get AIs to do what we want, and failing to, because we just don’t know how to specify what we want. And that’s, I think, basically the core of the AI safety problem — is that we don’t have a good way of specifying what we want. An example of that is what are called adversarial examples, which sort of demonstrate that computer vision systems that are able to do a really amazing job at classifying images and seeing what’s in an image and labeling images still make mistakes that humans just would never make. Images that look indistinguishable to humans can look completely different to the AI system, and that means that we haven’t really successfully communicated to the AI system what our visual concepts are. And so even though we think we have done a good job of telling it what to do, it’s like, “tell us what this picture is of” — the way that it found to do that really isn’t the way that we would do it and actually there’s some very problematic and unsettling differences there. And that’s another field that, along with the ones that I mentioned, like generative models and GANs, has been receiving a lot more attention in the last couple of years, which is really exciting from the point of view of safety and specification.</p><p><strong>Ariel:</strong> So, would it be fair to say that you think we’ve had progress or at least seen progress in addressing long-term safety issues, but some of the near-term safety issues, maybe we need faster work?</p><p><strong>David:</strong> I mean I think to be clear, we have such a long way to go to address the kind of issues we’re going to see with generally intelligent and super intelligent AIs, that I still think that’s an even more pressing problem, and that’s what I’m personally focused on. I just think that you can see that there are going to be a lot of really big problems in the near term as well. And we’re not even well equipped to deal with those problems right now.</p><p><strong>Roman:</strong> I’ll generally agree with David. I’m more concerned about long-term impacts. There are both more challenging and more impactful. It seems like short-term things may be problematic right now, but the main difficulty is that we didn’t start working on them in time. So problems like algorithmic fairness, bias, technological unemployment, are social issues which are quite solvable; They are not really that difficult from engineering or technical points of view. Whereas long-term control of systems which are more intelligent than you are — very much unsolved at this point in any even toy model. So I would agree with the part about bigger concerns but I think current problems we have today, they are already impacting people, but the good news is we know how to do better.</p><p><strong>David:</strong> I’m not sure that we know how to do better exactly. Like I think a lot of these problems, it’s more of a problem of willpower and developing political solutions, so the ones that you mentioned. But with the deep fakes, this is something that I think requires a little bit more of a technical solution in the sense of how we organize our society so that people are either educated enough to understand this stuff, or so that people actually have someone they trust and have a reason to trust, who they can take their word for it on that.</p><p><strong>Roman:</strong> That sounds like a great job, I’ll take it.</p><p><strong>Ariel:</strong> It almost sounds like something we need to have someone doing in person, though.</p><p>So going back to this past year: were there, say, groups that formed, or research teams that came together, or just general efforts that, while maybe they didn’t produce something yet, you think could produce something good, either in safety or AI in general?</p><p><strong>David:</strong> I think something interesting is happening in terms of the way AI safety is perceived and talked about in the broader AI and machine learning community. It’s a little bit like this phenomenon where once we solve something people don’t consider it AI anymore. So I think machine learning researchers, once they actually recognize the problem that the safety community has been sort of harping on and talking about and saying like, “Oh, this is a big problem” — once they say, “Oh yeah, I’m working on this kind of problem, and that seems relevant to me,” then they don’t really think that it’s AI safety, and they’re like, “This is just part of what I’m doing, making something that actually generalizes well and learns the right concept, or making something that is actually robust, or being able to interpret the model that I’m building, and actually know how it works.”</p><p>These are all things that people are doing a lot of work on these days in machine learning that I consider really relevant for AI safety. So I think that’s like a really encouraging sign, in a way, that the community is sort of starting to recognize a lot of the problems, or at least instances of a lot of the problems that are going to be really critical for aligning generally intelligent AIs.</p><p><strong>Ariel:</strong> And Roman, what about you? Did you see anything sort of forming in the last year that maybe doesn’t have some specific result, but that seemed hopeful to you?</p><p><strong>Roman:</strong> Absolutely. So I’ve mentioned that there is very few actual AI safety researchers as compared to the number of AI developers, researchers directly creating more capable machines. But the growth rate is much better I think. The number of organizations, the number of people who show interest in it, the number of papers I think is growing at a much faster rate, and it’s encouraging because as David said, it’s kind of like this convergence if you will, where more and more people realize, “I cannot say I built an intelligent system if it kills everyone.” That’s just not what an intelligent system is.</p><p>So safety and security become integral parts of it. I think Stuart Russell has a great example where he talks about bridge engineering. We don’t talk about safe bridges and secure bridges — there’s just bridges. If it falls down, it’s not a bridge. Exactly the same is starting to happen here: People realize, “My system cannot fail and embarrass the company, I have to make sure it will not cause an accident.”</p><p><strong>David:</strong> I think that a lot of people are thinking about that way more and more, which is great, but there is a sort of research mindset, where people just want to understand intelligence, and solve intelligence. And I think that’s kind of a different pursuit. Solving intelligence doesn’t mean that you make something that is safe and secure, it just means you make something that’s really intelligent, and I would like it if people who had that mindset were still, I guess, interested in or respectful of or recognized that this research is potentially dangerous. I mean, not right now necessarily, but going forward I think we’re going to need to have people sort of agree on having that attitude to some extent of being careful.</p><p><strong>Ariel:</strong> Would you agree though that you’re seeing more of that happening?</p><p><strong>David:</strong> Yeah, absolutely, yeah. But I mean it might just happen naturally on its own, which would be great.</p><p><strong>Ariel:</strong> Alright, so before I get to my very last question, is there anything else you guys wanted to bring up about 2018 that we didn’t get to yet?</p><p><strong>David:</strong> So we were talking about AI safety and there’s kind of a few big developments in the last year. I mean, there’s actually too many I think for me to go over all of them, but I wanted to talk about something which I think is relevant to the specification problem that I was talking about earlier.</p><p><strong>Ariel:</strong> Okay.</p><p><strong>David:</strong> So, there are three papers in the last year, actually, on what I call superhuman feedback. The idea motivating these works is that even specifying what we want on a particular instance in some particular scenario can be difficult. So typically the way that we would think about training an AI that understands our intentions is to give it a bunch of examples, and say, “In this situation, I prefer if you do this. This is the kind of behavior I want,” and then the AI is supposed to pick up on the patterns there and sort of infer what our intentions are more generally.</p><p>But there can be some things that we would like AI systems to be competent at doing, ideally, that are really difficult to even assess individual instances of. Two examples that I like to use are designing a transit system for a large city, or maybe for a whole country, or the world or something. That’s something that right now is done by a massive team of people. Using that whole team to sort of assess a proposed design that the AI might make would be one example of superhuman feedback, because it’s not just a single human. But you might want to be able to do this with just a single human and a team of AIs helping them, instead of a team of humans. And there’s a few proposals for how you could do that that have come out of the safety community recently, which I think are pretty interesting.</p><p><strong>Ariel:</strong> Why is it called superhuman feedback?</p><p><strong>David:</strong> Actually, this is just my term for it. I don’t think anyone else is using this term.</p><p><strong>Ariel:</strong> Okay.</p><p><strong>David:</strong> Sorry if that wasn’t clear. The reason I use it is because there are three different, like, lines of work here. So there’s these two papers from OpenAI on what’s called amplification and debate, and then another paper from DeepMind on reward learning and recursive reward learning. And I like to view these as all kind of trying to solve the same problem. How can we assist humans and enable them to make good judgements and informed judgements that actually reflect what their preferences are when they’re not capable of doing that by themselves unaided. So it’s superhuman in the sense that it’s better than a single human can do. And these proposals are also aspiring to do things I think that even teams of humans couldn’t do by having AI helpers that sort of help you do the evaluation.</p><p>An example that Yan — who’s the lead author on the DeepMind paper, which I also worked on — gives is assessing an academic paper. So if you yourself aren’t familiar with the field and don’t have the expertise to assess this paper, you might not be able to say whether or not it should be published. But if you can decompose that task into things like: is the paper valid? Are the proofs valid? Are the experiments following a reasonable protocol? Is it novel? Is it formatted correctly for the venue where it’s submitted? And you got answers to all of those from helpers, then you could make the judgment. You’d just be like okay, it meets all of the criteria, so it should be published. The idea would be to get AI helpers to do those sorts of evaluations for you across a broad range of tasks, and allow us to explain to AIs, or teach AIs what we want across a broad range of tasks in that way.</p><p><strong>Ariel:</strong> So, okay, and so then were there other things that you wanted to mention as well?</p><p><strong>David:</strong> I do feel like I should talk about another thing that was, again, not developed last year, but really sort of took off last year — is this new kind of neural network architecture called the transformer, which is basically being used in a lot of places where convolutional neural networks and recurrent neural networks were being used before. And those were kind of the two main driving factors behind the deep learning revolution in terms of vision, where you use convolutional networks and things that have a sequential structure, like speech, or text, where people were using recurrent neural networks. And this architecture is actually motivated originally by the same sort of scaling consideration because it allowed them to remove some of the most computationally heavy parts of running these kind of models in the context of translation, and basically make it a hundred times cheaper to train a translation model. But since then it’s also been used in a lot of other contexts and has shown to be a really good replacement for these other kinds of models for a lot of applications.</p><p>And I guess the way to describe what it’s doing is it’s based on what’s called an attention mechanism, which is basically a way of giving a neural network the ability to pay more attention to different parts of an input than other parts. So like to look at one word that is most relevant to the current translation task. So if you’re imagining outputting words one at a time, then because different languages have words in different order, it doesn’t make sense to sort of try and translate the next word. You want to look through the whole input sentence, like a sentence in English, and find the word that corresponds to whatever word should come next in your output sentence.</p><p>And that was sort of the original inspiration for this attention mechanism, but since then it’s been applied in a bunch of different ways, including paying attention to different parts of the model’s own computation, paying attention to different parts of images. And basically just using this attention mechanism in the place of the other sort of neural architectures that people thought were really important to give you temporal dependencies across something sequential like a sentence that you’re trying to translate, turned out to work really well.</p><p><strong>Ariel:</strong> So I want to actually pass this to Roman real quick. Did you have any comments that you wanted to add to either the superhuman feedback or the transformer architecture?</p><p><strong>Roman:</strong> Sure, so superhuman feedback: I like the idea and I think people should be exploring that, but we can kind of look at similar examples previously. So, for a while we had situation where teams of human chess players and machines did better than just unaided machines or unaided humans. That lasted about ten years. And then machines became so much better, humans didn’t really contribute anything, it was kind of just like an additional bottleneck to consult with them. I wonder if long term this solution will face similar problems. It’s very useful right now, but it seems like, I don’t know if it will scale.</p><p><strong>David:</strong> Well I want to respond to that, because I think it’s — the idea here is, in my mind, to have something that actually scales in the way that you’re describing, where it can sort of out-compete pure AI systems. Although I guess some people might be hoping that that’s the case, because that would make the strategic picture better in terms of people’s willingness to use safer systems. But this is more about just how can we even train systems — if we have the willpower, if people want to build a system that has the human in charge, and ends up doing what the human wants — how can we actually do that for something that’s really complicated?</p><p><strong>Roman:</strong> Right. And as I said, I think it’s a great way to get there. So this part I’m not concerned about. It’s a long-term game with that.</p><p><strong>David:</strong> Yeah, no, I mean I agree that that is something to be worried about as well.</p><p><strong>Roman:</strong> There is a possibility of manipulation if you have a human in the loop, and that itself makes it not safer but more dangerous in certain ways.</p><p><strong>David:</strong> Yeah, one of the biggest concerns I have for this whole line of work is that the human needs to really trust the AI systems that are assisting it, and I just don’t see that we have good enough mechanisms for establishing trust and building trustworthy systems right now, to really make this scale well without introducing a lot of risk for things like manipulation, or even just compounding of errors.</p><p><strong>Roman:</strong> But those approaches, like the debate approach, it just feels like they’re setting up humans for manipulation from both sides, and who’s better at breaking the human psychological model.</p><p><strong>David:</strong> Yep, I think it’s interesting, and I think it’s a good line of work. But I think we haven’t seen anything that looks like a convincing solution to me yet.</p><p><strong>Roman:</strong> Agreed.</p><p><strong>Ariel:</strong> So, Roman, was there anything else that you wanted to add about things that happened in the last year that we didn’t get to?</p><p><strong>Roman:</strong> Well, as a professor, I can tell you that students stop learning after about 40 minutes. So I think at this point we’re just being counterproductive.</p><p><strong>Ariel:</strong> So for what it’s worth, our most popular podcasts have all exceeded two hours. So, what are you looking forward to in 2019?</p><p><strong>Roman:</strong> Are you asking about safety or development?</p><p><strong>Ariel:</strong> Whatever you want to answer. Just sort of in general, as you look toward 2019, what relative to AI are you most excited and hopeful to see, or what do you predict we’ll see?</p><p><strong>David:</strong> So I’m super excited for people to hopefully pick up on this reward learning agenda that I mentioned that Jan and me and people at DeepMind worked on. I was actually pretty surprised how little work has been done on this. So the idea of this agenda at a high level is just: we want to learn a reward function — which is like a score, that tells an agent how well it’s doing — learn reward functions that encode what we want the AI to do, and that’s the way that we’re going to specify tasks to an AI. And I think from a machine learning researcher point of view this is kind of the most obvious solution to specification problems and to safety — is just learner reward function. But very few people are really trying to do that, and I’m hoping that we’ll see more people trying to do that, and encountering and addressing some of the challenges that come up.</p><p><strong>Roman:</strong> So I think by definition we cannot predict short-term breakthroughs. So what we’ll see is a lot of continuation of 2018 work, and previous work scaling up. So, if you have, let’s say, Texas hold ’em poker: so for two players, we’ll take it to six players, ten players, something like that. And you can make similar projections for other fields, so the strategy games will be taken to new maps, involve more players, maybe additional handicaps will be introduced for the bots. But that’s all we can really predict, kind of gradual improvement.</p><p>Protein folding will be even more efficient in terms of predicting actual structures: Any type of accuracy rates, if they were climbing from 80% to 90%, will hit 95, 96. And this is a very useful way of predicting what we can anticipate, and I’m trying to do something similar with accidents. So if we can see historically what was going wrong with systems, we can project those trends forward. And I’m happy to say that there is now at least two or three different teams working and collecting those examples and trying to analyze them and create taxonomies for them. So that’s very encouraging.</p><p><strong>David:</strong> Another thing that comes to mind is — I mentioned adversarial examples earlier, which are these imperceptible differences to a human that change how the AI system perceives something like an image. And so far, for the most part, the field has been focused on really imperceptible changes. But I think now people are starting to move towards a broader idea of what counts as an adversarial example. So basically anything that a human thinks clearly should belong to this class and the AI system thinks clearly should belong to this other class that has sort have been constructed deliberately to create that kind of a difference.</p><p>And I think this going to be really interesting and exciting to see how the field tries to move in that direction, because as I mentioned, I think it’s hard to define how humans decide whether or not something is a picture of a cat or something. And the way that we’ve done it so far is just by giving lots of examples of things that we say are cats. But it turns out that that isn’t sufficient, and so I think this is really going to push a lot of people closer towards thinking about some of the really core safety challenges within the mainstream machine learning community. So I think that’s super exciting.</p><p><strong>Roman:</strong> It is a very interesting topic, and I am in particular looking at a side subject in that, which is adversarial inputs for humans, and machines developing which I guess is kind of like optical illusions, and audio illusions, where a human is mislabeling inputs in a predictable way, which is allowing for manipulation.</p><p><strong>Ariel:</strong> Along very similar lines, I think I want to modify my questions slightly, and also ask: coming up in 2019, what are you both working on that you’re excited about, if you can tell us?</p><p><strong>Roman:</strong> Sure, so there has been a number of publications looking at particular limitations, either through mathematical proofs or through well known economic models, and what is possible in fact, from computational, complexity points of view. And I’m trying to kind of integrate those into a single model showing — in principle, not in practice, but even in principle — what can we do with the AI control problem? How solvable is it? Is it solvable? Is it not solvable? Because I don’t think there is a mathematically rigorous proof, or even a rigorous argument either way. So I think that will be helpful, especially with kind of arguing about importance of a problem and resource allocation.</p><p><strong>David:</strong> I’m trying to think what I can talk about. I guess right now I have some ideas for projects that are not super well thought out, so I won’t talk about those. And I have a project that I’m trying to finish off which is a little bit hard to describe in detail, but I’ll give the really high level motivation for it. And it’s about something that people in the safety community like to call capability control. I think Nick Bostrom has these terms, capability control and motivation control. And so what I’ve been talking about most of the time in terms of safety during this podcast was more like motivation control, like getting the AI to want to do the right thing, and to understand what we want. But that might end up being too hard, or sort of limited in some respect. And the alternative is just to make AIs that aren’t capable of doing things that are dangerous or catastrophic.</p><p>A lot of people in the safety community sort of worry about capability control approaches failing because if you have a very intelligent agent, it will view these attempts to control it as undesirable, and try and free itself from any constraints that we give it. And I think a way of sort of trying to get around that problem is to sort of look at capability control from the lens of motivation control. So to basically make an AI that doesn’t want to influence certain things, and maybe doesn’t have some of these drives to influence the world, or to influence the future. And so in particular I’m trying to see how can we design agents that really don’t try to influence the future, and really only care about doing the right thing, right now. And if we try and do that in a sort of naïve way, or there ways that can fail, and we can get some sort of emergent drive to still try and optimize over the long term, or try and have some influence in the future. And I think to the extent we see things like that, that’s problematic from this perspective of let’s just make AIs that aren’t capable or motivated to influence the future.</p><p><strong>Ariel:</strong> Alright! I think I’ve kept you both on for quite a while now. So, David and Roman, thank you so much for joining us today.</p><p><strong>David:</strong> Yeah, thank you both as well.</p><p><strong>Roman:</strong> Thank you so much.</p><p><em>Originally published at </em><a href="https://futureoflife.org/2019/01/31/fli-podcast-ai-breakthroughs-and-challenges-in-2018-with-david-krueger-and-roman-yampolskiy/"><em>futureoflife.org</em></a><em> on January 31, 2019.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b49fcdb0f20a" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI Alignment Podcast: The Byzantine Generals’ Problem, Poisoning, and Distributed Machine Learning…]]></title>
            <link>https://medium.com/@FLIxrisk/ai-alignment-podcast-the-byzantine-generals-problem-poisoning-and-distributed-machine-learning-ccd71c18e388?source=rss-e33e2d2a809c------2</link>
            <guid isPermaLink="false">https://medium.com/p/ccd71c18e388</guid>
            <category><![CDATA[machine-learning]]></category>
            <dc:creator><![CDATA[Future of Life]]></dc:creator>
            <pubDate>Wed, 30 Jan 2019 18:55:40 GMT</pubDate>
            <atom:updated>2019-02-07T20:45:32.443Z</atom:updated>
            <content:encoded><![CDATA[<h3>AI Alignment Podcast: The Byzantine Generals’ Problem, Poisoning, and Distributed Machine Learning with El Mahdi El Mhamdi (Beneficial AGI 2019)</h3><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fw.soundcloud.com%2Fplayer%2F%3Furl%3Dhttps%253A%252F%252Fapi.soundcloud.com%252Ftracks%252F571389006%26show_artwork%3Dtrue&amp;url=https%3A%2F%2Fsoundcloud.com%2Ffutureoflife%2Fthe-byzantine-generals-problem-poisoning-and-distributed-machine-learning-with-el-mahdi-el-mahmdi-beneficial-agi-2019&amp;image=http%3A%2F%2Fi1.sndcdn.com%2Fartworks-000484519200-a6g2oh-t500x500.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=soundcloud" width="800" height="166" frameborder="0" scrolling="no"><a href="https://medium.com/media/fb698d06d95ddc2c6283da71d4715850/href">https://medium.com/media/fb698d06d95ddc2c6283da71d4715850/href</a></iframe><p>Three generals are voting on whether to attack or retreat from their siege of a castle. One of the generals is corrupt and two of them are not. What happens when the corrupted general sends different answers to the other two generals?</p><p>A <a href="https://en.wikipedia.org/wiki/Byzantine_fault">Byzantine fault</a> is “<em>a condition of a computer system, particularly distributed computing systems, where components may fail and there is imperfect information on whether a component has failed. The term takes its name from an allegory, the “Byzantine Generals’ Problem”, developed to describe this condition, where actors must agree on a concerted strategy to avoid catastrophic system failure, but some of the actors are unreliable.“</em></p><p>The Byzantine Generals’ Problem and associated issues in maintaining reliable distributed computing networks is illuminating for both AI alignment and modern networks we interact with like Youtube, Facebook, or Google. By exploring this space, we are shown the limits of reliable distributed computing, the safety concerns and threats in this space, and the tradeoffs we will have to make for varying degrees of efficiency or safety.</p><p><em>The Byzantine Generals’ Problem, Poisoning, and Distributed Machine Learning with El Mahdi El Mhamdi</em> is the ninth podcast in the AI Alignment Podcast series, hosted by Lucas Perry. El Mahdi pioneered Byzantine resilient machine learning devising a series of provably safe algorithms he recently presented at NeurIPS and ICML. Interested in theoretical biology, his work also includes the analysis of error propagation and networks applied to both neural and biomolecular networks. This particular episode was recorded at the <a href="https://futureoflife.org/beneficial-agi-2019/">Beneficial AGI 2019</a> conference in Puerto Rico. We hope that you will join in the conversations by following us or subscribing to our podcasts on <a href="https://www.youtube.com/channel/UC-rCCy3FQ-GItDimSR9lhzw">Youtube</a>, <a href="https://soundcloud.com/futureoflife/tracks">SoundCloud</a>, <a href="https://itunes.apple.com/gb/podcast/the-future-of-life/id1170991978?mt=2">iTunes</a>, <a href="https://play.google.com/music/listen?u=0#/ps/Iatasczlcpeyawgv5werdrxbnh4">Google Play</a>, <a href="https://www.stitcher.com/podcast/future-of-life-institute/the-future-of-life-institue">Stitcher</a>, or your preferred podcast site/application. You can find all the <a href="https://futureoflife.org/ai-alignment-podcast/">AI Alignment Podcasts here</a>.</p><p>If you’re interested in exploring the interdisciplinary nature of AI alignment, we suggest you take a look <a href="https://futureoflife.org/valuealignmentmap/">here at a preliminary landscape</a> which begins to map this space.</p><p>Topics discussed in this episode include:</p><ul><li>The Byzantine Generals’ Problem</li><li>What this has to do with artificial intelligence and machine learning</li><li>Everyday situations where this is important</li><li>How systems and models are to update in the context of asynchrony</li><li>Why it’s hard to do Byzantine resilient distributed ML.</li><li>Why this is important for long-term AI alignment</li></ul><p>An overview of Adversarial Machine Learning and where Byzantine-resilient Machine Learning stands on the map is available in this <a href="https://www.youtube.com/watch?v=1Ds1DAnoaag/">(9min) video</a>. A specific focus on Byzantine Fault Tolerant Machine Learning is available <a href="https://www.youtube.com/watch?v=SaHBLGv_X2E/">here (~7min)</a>.</p><p>In particular, El Mahdi argues in the first interview (and in the podcast) that technical AI safety is not only relevant for long term concerns, but is crucial in current pressing issues such as social media poisoning of public debates and misinformation propagation, both of which fall into Poisoning-resilience. Another example he likes to use is social media addiction, that could be seen as a case of (non) Safely Interruptible learning. This value misalignment is already an issue with the primitive forms of AIs that optimize our world today as they maximize our watch-time all over the internet.</p><p>The latter (Safe Interruptibility) is another technical AI safety question El Mahdi works on, in the context of Reinforcement Learning. This line of research was initially dismissed as “science fiction”, <a href="https://www.youtube.com/watch?v=9O1fsvzMA6k//">in this interview (5min</a>), El Mahdi explains why it is a realistic question that arises naturally in reinforcement learning</p><p>“El Mahdi’s work on Byzantine-resilient Machine Learning and other relevant topics is available on his <a href="https://scholar.google.ch/citations?hl=de&amp;user=kNA-WLQAAAAJ&amp;view_op=list_works&amp;sortby=pubdate//">Google scholar profile</a>. A modification of the popular machine learning library TensorFlow, to make it Byzantine-resilient (and also support communication over UDP channels among other things) has been recently <a href="https://github.com/LPD-EPFL/AggregaThor">open-sourced on Github by El Mahdi’s</a> colleagues based on his algorithmic work we mention in the podcast.</p><p><a href="https://twitter.com/L_badikho/">To connect with him over social media</a></p><p>You can listen to the podcast above, or read the full transcript below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1eQI7jo-aN6hsg43NpqVXQ.gif" /></figure><p><strong>Lucas: </strong>Hey, everyone. Welcome back to the AI Alignment Podcast series. I’m Lucas Perry, and today we’ll be speaking with El Mahdi El Mhamdi on the Byzantine problem, Byzantine tolerance, and poisoning in distributed learning and computer networks. If you find this podcast interesting or useful, please give it a like and follow us on your preferred listing platform. El Mahdi El Mhamdi pioneered Byzantine resilient machine learning devising a series of provably safe algorithms he recently presented at NeurIPS and ICML. Interested in theoretical biology, his work also includes the analysis of error propagation and networks applied to both neural and biomolecular networks. With that, El Mahdi’s going to start us off with a thought experiment.</p><p><strong>El Mahdi:</strong> Imagine you are part of a group of three generals, say, from the Byzantine army surrounding a city you want to invade, but you also want to retreat if retreat is the safest choice for your army. You don’t want to attack when you will lose, so those three generals that you’re part of are in three sides of the city. They sent some intelligence inside the walls of the city, and depending on this intelligence information, they think they will have a good chance of winning and they would like to attack, or they think they will be defeated by the city, so it’s better for them to retreat. Your final decision would be a majority vote, so you communicate through some horsemen that, let’s say, are reliable for the sake of this discussion. But there might be one of you who might have been corrupt by the city.</p><p>The situation would be problematic if, say, there are General A, General B, and General C. General A decided to attack. General B decided to retreat based on their intelligence for some legitimate reason. A and B are not corrupt, and say that C is corrupt. Of course, A and B, they can’t figure out who was corrupt. Say C is corrupt. What this general would do they thing, so A wanted to attack. They will tell them, “I also want to attack. I will attack.” Then they will tell General B, “I also want to retreat. I will retreat.” A receives two attack votes and one retreat votes. General B receives two retreat votes and only one attack votes. If they trust everyone, they don’t do any double checking, this would be a disaster.</p><p>A will attack alone; B would retreat; C, of course, doesn’t care because he was corrupt by the cities. You can tell me they can circumvent that by double checking. For example, A and B can communicate on what C told them. Let’s say that every general communicates with every general on what he decides and on also what’s the remaining part of the group told them. A will report to B, “General C told me to attack.” Then B would tell C, “General C told me to retreat.” But then A and B wouldn’t have anyway of concluding whether the inconsistency is coming from the fact that C is corrupt or that the general reporting on what C told them is corrupt.</p><p>I am General A. I have all the valid reasons to think with the same likelihood that C is maybe lying to me or also B might also be lying to me. I can’t know if you are misreporting what C told you enough for the city to corrupt one general if there are three. It’s impossible to come up with an agreement in this situation. You can easily see that this will generalize to having more than three generals, like I say 100, as soon as the non-corrupt one are less than two-thirds because what we saw with three generals would happen with the fractions that are not corrupt. Say that you have strictly more than 33 generals out of 100 who are corrupt, so what they can do is they can switch the majority votes on each side.</p><p>But worse than that, say that you have 34 corrupt generals and the remaining 66 not corrupt generals. Say that those 66 not corrupt generals were 33 on the attack side, 33 on the retreat side. The problem is that when you are in some side, say that you are in the retreat side, you have in front of you a group of 34 plus 33 in which there’s a majority of malicious ones. This majority can collude. It’s part of the Byzantine hypothesis. The malicious ones can collude and they will report a majority of inconsistent messages on the minority on the 33 ones. You can’t provably realize that the inconsistency is coming from the group of 34 because they are a majority.</p><p><strong>Lucas:</strong> When we’re thinking about, say, 100 persons or 100 generals, why is it that they’re going to be partitioned automatically into these three groups? What if there’s more than three groups?</p><p><strong>El Mahdi:</strong> Here we’re doing the easiest form of Byzantine agreement. We want to agree on attack versus retreat. When it’s become multi-dimensional, it gets even messier. There are more impossibility results and impossibility results. Just like with the binary decision, there is an impossibility theorem on having agreement if you have unsigned messages to horsemen. Whenever the corrupt group exceeds 33%, you provably cannot come up with an agreement. There are many variants to this problem, of course, depending on what hypothesis you can assume. Here, without even mentioning it, we were assuming bounded delays. The horsemen would always arrive eventually. If the horsemen could die on the way and you don’t have any way to check whether they arrive or not or you can be waiting forever because you don’t have any proof that the horsemen died on the way.</p><p>You don’t have any mechanism to tell you, “Stop waiting for the horsemen. Stop waiting for the message from General B because the horsemen died.” You can be waiting forever and there are theorems that shows that when you have unbounded delays, and by the way, like in distributed computing, whenever you have in bounded delays, we speak about asynchrony. If you have a synchronous communication, there is a very famous theorem that tells you consensus is impossible, not even in the malicious case, but just like in …</p><p><strong>Lucas:</strong> In the mundane normal case.</p><p><strong>El Mahdi: </strong>Yes. It’s called the <a href="https://groups.csail.mit.edu/tds/papers/Lynch/jacm85.pdf">Fischer Lynch Patterson theorem theorem</a> .</p><p><strong>Lucas:</strong> Right, so just to dive down into the crux of the problem, the issue here fundamentally is that when groups of computers or groups of generals or whatever are trying to check who is lying amongst discrepancies and similarities of lists and everyone who’s claiming what is when there appears to be a simple majority within that level of corrupted submissions, then, yeah, you’re screwed.</p><p><strong>El Mahdi:</strong> Yes. It’s impossible to achieve agreement. There are always fractions of malicious agents above which is provably impossible to agree. Depending on the situation, it will be a third or sometimes or a half or a quarter, depending on your specifications.</p><p><strong>Lucas:</strong> If you start tweaking the assumptions behind the thought experiment, then it changes what number of corrupted machines or agents that are required in order to flip the majority and to poison the communication.</p><p><strong>El Mahdi:</strong> Exactly. But for example, you mentioned something very relevant to today’s discussion, which is what if we were not agreeing on two decisions, retreat, attack. What if we were agreeing on some multi-dimensional decision? Attack or retreat on one dimension and then …</p><p><strong>Lucas:</strong> Maybe hold, keep the siege going.</p><p><strong>El Mahdi: </strong>Yeah, just like add possibilities or dimensions and multi-dimensional agreements. They’re even more hopeless results in that direction</p><p><strong>Lucas:</strong> There are more like impossibility theorems and issues where these distributed systems are vulnerable to small amounts of systems being corrupt and screwing over the entire distributed network.</p><p><strong>El Mahdi:</strong> Yes. Maybe now we can slightly move to machine learning.</p><p><strong>Lucas: </strong>I’m happy to move into machine learning now. We’ve talked about this, and I think our audience can probably tell how this has to do with computers. Yeah, just dive in what this has to do with machine learning and AI and current systems today, and why it even matters for AI alignment.</p><p><strong>El Mahdi:</strong> As a brief transition, solving the agreement problem besides this very nice historic thought experiment is behind consistencies of safety critical systems like banking systems. Imagine we have a shared account. Maybe you remove 10% of the amount and then she or he added some $10 to the accounts. You remove the $10 in New York and she or he put the $10 in Los Angeles. The banking system has to agree on the ordering because minus $10 plus 10% is not the same result as plus 10% then minus $10. The final balance of the account would not be the same.</p><p><strong>Lucas:</strong> Right.</p><p><strong>El Mahdi:</strong> The banking systems routinely are solving decisions that fall into agreement. If you work on some document sharing platform, like Dropbox or Google Docs, whatever, and we collaboratively are writing the document, me and you. The document sharing platform has to, on real time, solve agreements about the ordering of operations so that me and you always keep seeing the same thing. This has to happen while some of the machines that are interconnecting us are failing, whether just like failing because there was a electric crash or something. Data center has lost some machines or if it was like restart, a bug or a take away. What we want in distributed computing is that we would like communications schemes between machines that’s guarantee this consistency that comes from agreement as long as some fraction of machines are reliable. What this has to do with artificial intelligence and machine learning reliability is that with some colleagues, we are trying to encompass one of the major issues in machine learning reliability inside the Byzantine fault tolerance umbrella. For example, you take, for instance, poisoning attacks.</p><p><strong>Lucas: </strong>Unpack what poisoning attacks are.</p><p><strong>El Mahdi:</strong> For example, imagine you are training a model on what are good videos to recommend given some key word search. If you search for “medical advice for young parents on vaccine,” this is a label. Let’s assume for the sake of simplicity that a video that tells you not to take your kid for vaccines is not what we mean by medical advice for young parents on vaccine because that’s what medical experts agree on. We want our system to learn that anitvaxers, like anti-vaccine propaganda is not what people are searching for when they type those key words, so I suppose a world where we care about accuracy, okay? Imagine you want to train a machine learning model that gives you accurate results of your search. Let’s also for the sake of simplicity assume that a majority of people on the internet are honest.</p><p>Let’s assume that more than 50% of people are not actively trying to poison the internet. Yeah, this is very optimistic, but let’s assume that. What we can show and what me and my colleagues started this line of research with is that you can easily prove that one single malicious agent can provably poison a distributed machine learning scheme. Imagine you are this video sharing platform. Whenever people behave on your platform, this generates what we call gradients, so it updates your model. It only takes a few hyperactive accounts that could generate behavior that is powerful enough to pull what we call the average gradient because what distributed machine learning is using, at least up to today, if you read the source code of most distributed machine learning frameworks. Distributed machine learning is always averaging gradients.</p><p>Imagine you Lucas Perry just googled a video on the Parkland shootings. Then the video sharing platform shows you a video telling you that David Hogg and Emma Gonzalez and those kids behind the March for Our Lives movement are crisis actors. The video labels three kids as crisis actors. It obviously has a wrong label, so it is what I will call a poisoned data point. If you are non-malicious agents on the video sharing platform, you will dislike the video. You will not approve it. You’re likely to flag it. This should generate a gradient that pushes the model in that direction, so the gradient will update the model into a direction where it stops thinking that this video is relevant for someone searching “Parkland shooting survivors.” What can happen if your machine learning framework is just averaging gradients is that a bunch of hyperactive people on some topic could poison the average and pull it towards the direction where the models is enforcing this thinking that, “Yeah, those kids are crisis actors.”</p><p><strong>Lucas: </strong>This is the case because the hyperactive accounts are seen to be given more weight than accounts which are less active in the same space. But this extra weighting that these accounts will get from their hyperactivity in one certain category or space over another, how is the weighting done? Is it just time spent per category or does it have to do with submissions that agree with the majority?</p><p><strong>El Mahdi: </strong>We don’t even need to go into the details because we don’t know. I’m talking in a general setting where you have a video sharing platform aggregating gradients for behavior. Now, maybe let’s raise the abstraction level. You are doing gradient descents, so you have a lost function that you want to minimize. You have an error function. The error function is the mismatch between what you predict and what the user tells you. The user tells you this is a wrong prediction, and then you move to the direction where the users stop telling you this is the wrong direction. You are doing great in this sense minimizing the lost function. User behaves, and with their behavior, you generate gradients.</p><p>What you do now in the state of the arts way of distributed machine learning is that you average all those gradients. Averaging is well known not to be resilient. If you have a room of poor academics earning a few thousand dollars and then a billionaire jumps in the room, if your algorithm reasons with averaging, it will think that this is a room of millionaires because the average salary would be a couple of hundred millions. But then million is very obvious to do when it comes to salaries and numbers scalers because you can rank them.</p><p><strong>Lucas:</strong> Right.</p><p><strong>El Mahdi: </strong>You rank numbers and then decide, “Okay, this is the ordering. This is the number that falls in the middle. This is the upper half. This is the lower half and this is the median.” When it becomes high dimensional, the median is a bit tricky. It has some computational issues. Then even if you compute what we call the geometric median, an attacker can still know how to leverage the fact that you’re only approximating it because there’s no closed formula. There’s no closed form to compute the median in that dimension. But worse than that, what we showed in one of our follow up works is because of the fact that machine learning is done in very, very, very high dimensions, you would have a curse of the dimensionality issue that makes it possible for attackers to sneak in without being spot as a way of the median.</p><p>It can still look like the median vector. I take benefits from the fact that those vectors, those gradients, are extremely high dimensional. I would look for all the disagreements. Let’s say you have a group of a couple hundred gradients, and I’m the only malicious one. I would look at the group of correct vectors all updating you somehow in the same direction within some variants. On average, they’re like what we call unbiased estimators of the gradient. When you take out the randomness, the expected value they will give you is the real gradient of the loss function. What I will do as a malicious worker is I will look at the way they are disagreeing slightly on each direction.</p><p>I will sum that. I will see that they disagree by this much on direction one. They disagree by this much on direction two. They disagree by this much, epsilon one, epsilon two, epsilon three. I would look for all these small disagreements they have on all the components.</p><p><strong>Lucas: </strong>Across all dimensions and high dimensional space. [crosstalk 00:16:35]</p><p><strong>El Mahdi: </strong>Then add that up. It will be my budget, my leeway, my margin to attack you on another direction.</p><p><strong>Lucas: </strong>I see.</p><p><strong>El Mahdi: </strong>What we proved is that you have to mix ideas from geometric median with ideas from the traditional component-wise median, and that those are completely different things. The geometric median is a way to find a median by just minimizing the sum of distances between what you look for and all the vectors that were proposed, and then the component-wise median will do a traditional job of ranking the coordinates. It looks at each coordinate, and then rank all the propositions, and then look for the proposition that lies in the middle. Once we proved enough follow up work is that, yeah, the geometric median idea is elegant. It can make you converge, but it can make you converge to something arbitrarily bad decided by the attacker. When you train complex models like neural nets, the landscape you optimize inside is not convex. It’s not like a bowl or a cup where you just follow the descending slope you would end up in the lowest point.</p><p><strong>Lucas:</strong> Right.</p><p><strong>El Mahdi: </strong>It’s like a multitude of bowls with different heights.</p><p><strong>Lucas: </strong>Right, so there’s tons of different local minima across the space.</p><p><strong>El Mahdi: </strong>Exactly. So in the first paper what we showed is that ideas that look like the geometric median are enough to just converge. You converge. You provably converge, but in the follow up what we realized, like something we were already aware of, but not enough in my opinion, is that there is this square root D, this curse of dimensionality that will arise when you compute high dimensional distances. That the attacker can leverage.</p><p>So in what we call the hidden vulnerability of distributed learning, you can have correct vectors, agreeing on one component. Imagine in your head some three axis system.</p><p>Let’s say that they are completely in agreement on axis three. But then in axis one, two, so in the plane formed by the axis one and axis two, they have a small disagreement.</p><p>What I will do as the malicious agent, is that I will leverage this small disagreement, and inject it in axis three. And this will make you go to a bit slightly modified direction. And instead of going to this very deep, very good minima, you will go into a local trap that is just close ahead.</p><p>And that comes from the fact that loss functions of interesting models are clearly like far from being convex. The models are highly dimensional, and the loss function is highly un-convex, and creates a lot of leeway.</p><p><strong>Lucas: </strong>It creates a lot of local minima spread throughout the space for you to attack the person into.</p><p><strong>El Mahdi: </strong>Yeah. So convergence is not enough. So we started this research direction by formulating the following question, what does it take to guarantee convergence?</p><p>And any scheme that aggregates gradients, and guarantee convergence is called Byzantine resilient. But then you can realize that in very high dimensions, and highly non-convex loss functions, is convergence enough? Would you just want to converge?</p><p>There are of course people arguing the deep learning models, like there’s this famous paper by Anna Choromanska, and Yann LeCun, and Gérard Ben Arous, about the landscape of neural nets, that basically say that, “Yeah, very deep local minimum of neural nets are some how as good.”</p><p>From an overly simplified point of view, it’s an optimistic paper, that tells you that you shouldn’t worry too much when you optimize neural nets about the fact that gradient descent would not necessarily go to a global like-</p><p><strong>Lucas:</strong> To a global minima.</p><p><strong>El Mahdi: </strong>Yeah. Just like, “Stop caring about that.”</p><p><strong>Lucas: </strong>Because the local minima are good enough for some reason.</p><p><strong>El Mahdi: </strong>Yeah. I think that’s a not too unfair way to summarize the paper for the sake of this talk, for the sake of this discussion. What we empirically illustrate here, and theoretically support is that that’s not necessarily true.</p><p>Because we show that with very low dimensional, not extremely complex models, trained on CIFAR-10 and MNIST, which are toy problems, very easy toy problems, low dimensional models etc. It’s already enough to have those amounts of parameters, let’s say 100,000 parameters or less, so that an attacker would always find a direction to take you each time away, away, away, and then eventually find an arbitrarily bad local minimum. And then you just converge to that.</p><p>So convergence is not enough. Not only you have to seek an aggregation rule that guarantees convergence, but you have to seek some aggregation rules that guarantee that you would not converge to something arbitrarily bad. You would keep converging to the same high quality local minimum, whatever that means.</p><p>The hidden vulnerability is this high dimensional idea. It’s the fact that because the loss function is highly non-convex, because there’s the high dimensionality, as an attacker I would always find some direction, so the attack goes this way.</p><p>Here the threat model is that an attacker can spy on your gradients, generated by the correct workers but cannot talk on their behalf. So I cannot corrupt the messages. Since you asked about the reliability of horsemen or not.</p><p>So horsemen are reliable. I can’t talk on your behalf, but I can spy on you. I can see what are you sending to the others, and anticipate.</p><p>So I would as an attacker wait for correct workers to generate their gradients, I will gather those vectors, and then I will just do a linear regression on those vectors to find the best direction to leverage the disagreement on the D minus one remaining directions.</p><p>So because there would be this natural disagreement, this variance in many directions, I will just do some linear regression and find what is the best direction to keep? And use the budget I gathered, those epsilons I mentioned earlier, like this D time epsilon on all the directions to inject it the direction that will maximize my chances of taking you away from local minima.</p><p>So you will converge, as proven in the early papers, but not necessarily to something good. But what we showed here is that if you combine ideas from multidimensional geometric medians, with ideas from single dimensional component-wise median, you improve your robustness.</p><p>Of course it comes with a price. You require three quarters of the workers to be reliable.</p><p>There is another direction where we expanded this problem, which is asynchrony. And asynchrony arises when as I said in the Byzantine generals setting, you don’t have a bounded delay. In the bounded delay setting, you know that horses arrive at most after one hour.</p><p><strong>Lucas: </strong>But I have no idea if the computer on the other side of the planet is ever gonna send me that next update.</p><p><strong>El Mahdi: </strong>Exactly. So imagine you are doing machine learning on smartphones. You are leveraging a set of smartphones all around the globe, and in different bandwidths, and different communication issues etc.</p><p>And you don’t want each time to be bottlenecked by the slowest one. So you want to be asynchronous, you don’t want to wait. You’re just like whenever some update is coming, take it into account.</p><p>Imagine some very advanced AI scenario, where you send a lot of learners all across the universe, and then they communicate with the speed of light, but some of them are five light minutes away, but some others are two hours and a half. And you want to learn from all of them, but not necessarily handicap the closest one, because there are some other learners far away.</p><p><strong>Lucas: </strong>You want to run updates in the context of asynchrony.</p><p><strong>El Mahdi:</strong> Yes. So you want to update whenever a gradient is popping up.</p><p><strong>Lucas: </strong>Right. Before we move on to illustrate the problem again here is that the order matters, right? Like in the banking example. Because the 10% plus 10 is different from-</p><p><strong>El Mahdi: </strong>Yeah. Here the order matters for different reasons. You update me so you are updating me on the model you got three hours ago. But in the meanwhile, three different agents updated me on the models, while getting it three minutes ago.</p><p>All the agents are communicating through some abstraction they call the server maybe. Like this server receives updates from fast workers.</p><p><strong>Lucas: </strong>It receives gradients.</p><p><strong>El Mahdi: </strong>Yeah, gradients. I also call them updates.</p><p><strong>Lucas:</strong> Okay.</p><p><strong>El Mahdi:</strong> Because some workers are close to me and very fast, I’ve done maybe 1000 updates, while you were still working and sending me the message.</p><p>So when your update arrive, I can tell whether it is very stale, very late, or malicious. So what we do in here is that, and I think it’s very important now to connect a bit back with classic distributed computing.</p><p>Is that Byzantine resilience in machine learning is easier than Byzantine resilience in classical distributed computing for one reason, but it is extremely harder for another reason.</p><p>The reason is that we know what we want to agree on. We want to agree on a gradient. We have a toolbox of calculus that tells us how this looks like. We know that it’s the slope of some loss function that is most of today’s models, relatively smooth, differentiable, maybe Lipschitz, bounded, whatever curvature.</p><p>So we know that we are agreeing on vectors that are gradients of some loss function. And we know that there is a majority of workers that will produce vectors that will tell us what does a legit vector look like.</p><p>You can find some median behavior, and then come up with filtering criterias that will get away with the bad gradients. That’s the good news. That’s why it’s easier to do Byzantine resilience in machine learning than to do Byzantine agreement. Byzantine agreement, because agreement is a way harder problem.</p><p>The reason why Byzantine resilience is harder in machine learning than in the typical settings you have in distributed computing is that we are dealing with extremely high dimensional data, extremely high dimensional decisions.</p><p>So a decision here is to update the model. It is triggered by a gradient. So whenever I accept a gradient, I make a decision. I make a decision to change the model, to take it away from this state, to this new state, by this much.</p><p>But this is a multidimensional update. And Byzantine agreement, or Byzantine approximate agreement in higher dimension has been provably hopeless by Hammurabi Mendes, and Maurice Herlihy in an excellent paper in 2013, where they show that you can’t do Byzantine agreement in D dimension with N agents in less than N to the power D computations, per agent locally.</p><p>Of course in their paper, they were meaning Byzantine agreement on positions. So they were framing it with a motivations saying, “This is N to the power D, but the typical cases we care about in distributed computing are like robots agreeing on a position on a plane, or on a position in a three dimensional space.” So D is two or three.</p><p>So N to the power two or N to the power three is fine. But in machine learning D is not two and three, D is a billion or a couple of millions. So N to the power a million is just like, just forget.</p><p>And not only that, but also they require … Remember when I tell you that Byzantine resilience computing would always have some upper bound on the number malicious agents?</p><p><strong>Lucas: </strong>Mm-hmm (affirmative).</p><p><strong>El Mahdi: </strong>So the number of total agents should exceed D times the number of malicious agents.</p><p><strong>Lucas: </strong>What is D again sorry?</p><p><strong>El Mahdi: </strong>Dimension.</p><p><strong>Lucas: </strong>The dimension. Okay.</p><p><strong>El Mahdi: </strong>So if you have to agree on D dimension, like on a billion dimensional decision, you need at least a billion times the number of malicious agents.</p><p>So if you have say 100 malicious agents, you need at least 100 billion total number of agents to be resistant. No one is doing distributed machine learning on 100 billion-</p><p><strong>Lucas: </strong>And this is because the dimensionality is really screwing with the-</p><p><strong>El Mahdi: </strong>Yes. Byzantine approximate agreement has been provably hopeless. That’s the bad, that’s why the dimensionality of machine learning makes it really important to go away, to completely go away from traditional distributed computing solutions.</p><p><strong>Lucas: </strong>Okay.</p><p><strong>El Mahdi: </strong>So we are not doing agreement. We’re not doing agreement, we’re not even doing approximate agreement. We’re doing something-</p><p><strong>Lucas: </strong>Totally new.</p><p><strong>El Mahdi: </strong>Not new, totally different.</p><p><strong>Lucas: </strong>Okay.</p><p><strong>El Mahdi: </strong>Called gradient decent. It’s not new. It’s as old as Newton. And it comes with good news. It comes with the fact that there are some properties, like some regularity of the loss function, some properties we can exploit.</p><p>And so in the asynchronous setting, it becomes even more critical to leverage those differentiability properties. So because we know that we are optimizing a loss functions that has some regularities, we can have some good news.</p><p>And the good news has to do with curvature. What we do here in asynchronous setting, is not only we ask workers for their gradients, we ask them for their empirical estimate of the curvature.</p><p><strong>Lucas: </strong>Sorry. They’re estimating the curvature of the loss function, that they’re adding the gradient to?</p><p><strong>El Mahdi:</strong> They add the gradient to the parameter, not the loss function. So we have a loss function, parameter is the abscissa, you add the gradient to the abscissa to update the model, and then you end up in a different place of the loss function.</p><p>So you have to imagine the loss function as like a surface, and then the parameter space as the plane, the horizontal plane below the surface. And depending on where you are in the space parameter, you would be on different heights of the loss function.</p><p><strong>Lucas: </strong>Wait sorry, so does the gradient depend where you are on this, the bottom plane?</p><p><strong>El Mahdi: </strong>Yeah [crosstalk 00:29:51]-</p><p><strong>Lucas:</strong> So then you send an estimate for what you think the slope of the intersection will be?</p><p><strong>El Mahdi:</strong> Yeah. But for asynchrony, not only that. I will ask you to send me the slope, and your observed empirical growth of the slope.</p><p><strong>Lucas: </strong>The second derivative?</p><p><strong>El Mahdi: </strong>Yeah.</p><p><strong>Lucas: </strong>Okay.</p><p><strong>El Mahdi:</strong> But the second derivative again in high dimension is very hard to compute. You have to computer the Hessian matrix.</p><p><strong>Lucas: </strong>Okay.</p><p><strong>El Mahdi: </strong>That’s something like completely ugly to compute in high dimensional situations because it takes D square computations.</p><p>As an alternative we would like you to send us some linear computation in D, not a square computation in D.</p><p>So we would ask you to compute your actual gradient, your previous gradient, the difference between them, and normalize it by the difference between models.</p><p>So, “Tell us your current gradient, by how much it changed from the last gradient, and divide that by how much you changed the parameter.”</p><p>So you would tell us, “Okay, this is my current slope, and okay this is the gradient.” And you will also tell us, “By the way, my slope change relative to my parameter change is this much.”</p><p>And this would be some empirical estimation of the curvature. So if you are in a very curved area-</p><p><strong>Lucas: </strong>Then the estimation isn’t gonna be accurate because the linearity is gonna cut through some of the curvature.</p><p><strong>El Mahdi: </strong>Yeah but if you are in a very curved area of the loss function, your slope will change a lot.</p><p><strong>Lucas: </strong>Okay. Exponentially changing the slope.</p><p><strong>El Mahdi: </strong>Yeah. Because you did a very tiny change in the parameter and it takes a lot of the slope.</p><p><strong>Lucas: </strong>Yeah. Will change the … Yeah.</p><p><strong>El Mahdi: </strong>When you are in a non-curved area of the loss function, it’s less harmful for us that you are stale, because you will just technically have the same updates.</p><p>If you are in a very curved area of the loss function, your updates being stale is now a big problem. So we want to discard your updates proportionally to your curvature.</p><p>So this is the main idea of this scheme in asynchrony, where we would ask workers about their gradient, and their empirical growth rates.</p><p>And then of course I don’t want to trust you on what you declare, because you can plan to screw me with some gradients, and then declare a legitimate value of the curvature.</p><p>I will take those empirical, what we call in the paper empirical Lipschitz-ness. So we ask you for this empirical growth rate, that it’s a scalar, remember? This is very important. It’s a single dimensional number.</p><p>And so we ask you about this growth rate, and we ask all of you about growth rates, again assuming the majority is correct. So the majority of growth rates will help us set the median growth rate in a robust manner, because as long as a simple majority is not lying, the median growth rates will always be bounded between two legitimate values of the growth rate.</p><p><strong>Lucas: </strong>Right because, are you having multiple workers inform you of the same part of your loss function?</p><p><strong>El Mahdi: </strong>Yes. Even though they do it in an asynchronous manner.</p><p><strong>Lucas: </strong>Yeah. Then you take the median of all of them.</p><p><strong>El Mahdi:</strong> Yes. And then we reason by quantiles of the growth rates.</p><p><strong>Lucas: </strong>Reason by quantiles? What are quantiles?</p><p><strong>El Mahdi: </strong>The first third, the second third, the third third. Like the first 30%, the second 30%, the third 30%. We will discard the first 30%, discard the last 30%. Anything in the second 30% is safe.</p><p>Of course this has some level of pessimism, which is good for safety, but not very good for being fast. Because maybe people are not lying, so maybe the first 30%, and the last 30% are also values we could consider. But for safety reasons we want to be sure.</p><p><strong>Lucas:</strong> You want to try to get rid of the outliers.</p><p><strong>El Mahdi: </strong>Possible.</p><p><strong>Lucas:</strong> Possible outliers.</p><p><strong>El Mahdi:</strong> Yeah. So we get rid of the first 30%, the last 30%.</p><p><strong>Lucas: </strong>So this ends up being a more conservative estimate of the loss function?</p><p><strong>El Mahdi:</strong> Yes. That’s completely right. We explain that in the paper.</p><p><strong>Lucas:</strong> So there’s a trade off that you can decide-</p><p><strong>El Mahdi: </strong>Yeah.</p><p><strong>Lucas:</strong> By choosing what percentiles to throw away.</p><p><strong>El Mahdi: </strong>Yeah. Safety never comes for free. So here, depending on how good your estimates about the number of potential Byzantine actors is, your level of pessimism with translate into slowdown.</p><p><strong>Lucas: </strong>Right. And so you can update the amount that you’re cutting off-</p><p><strong>El Mahdi: </strong>Yeah.</p><p><strong>Lucas:</strong> Based off of the amount of expected corrupted signals you think you’re getting.</p><p><strong>El Mahdi: </strong>Yeah. So now imagine a situation where you know the number of workers is know. You know that you are leveraging 100,000 smartphones doing gradient descent for you. Let’s call that N.</p><p>You know that F of them might be malicious. We argue that if F is exceeding the third of N, you can’t do anything. So we are in a situation where F is less than a third. So less than 33,000 workers are malicious, then the slowdown would be F over N, so a third.</p><p>What if you are in a situation where you know that your malicious agents are way less than a third? For example you know that you have at most 20 rogue accounts in your video sharing platform.</p><p>And your video sharing platform has two billion accounts. So you have two billion accounts.</p><p><strong>Lucas: </strong>20 of them are malevolent.</p><p><strong>El Mahdi: </strong>What we show is that the slowdown would be N minus F divided by N. N is the two billion accounts, F is the 20, and is again two billion.</p><p>So it would be two billion minus 20, so one million nine hundred billion, like something like 0.999999. So you would go almost as fast as the non-Byzantine resilient scheme.</p><p>So our Byzantine resilient scheme has a slowdown that is very reasonable in situations where F, the number of malicious agents is way less than N, the total number of agents, which is typical in modern…</p><p>Today, like if you ask social media platforms, they have a lot of a tool kits to prevent people from creating a billion fake accounts. Like you can’t in 20 hours create an army of several million accounts.</p><p>None of the mainstream social media platforms today are susceptible to this-</p><p><strong>Lucas: </strong>Are susceptible to massive corruption.</p><p><strong>El Mahdi:</strong> Yeah. To this massive account creation. So you know that the number of corrupted accounts are negligible to the number of total accounts.</p><p>So that’s the good news. The good news is that you know that F is negligible to N. But then the slowdown of our Byzantine resilient methods is also close to one.</p><p>But it has the advantage compared to the state of the art today to train distributed settings of not taking the average gradient. And we argued in the very beginning that those 20 accounts that you could create, it doesn’t take a bot army or whatever, you don’t need to hack into the machines of the social network. You can have a dozen human, sitting somewhere in a house manually creating 20 accounts, training the accounts over time, doing behavior that makes the legitimate for some topics, and then because you’re distributing machine learning scheme would average the gradients generated by people behavior and that making your command anti-vaccine or controversies, anti-Semitic conspiracy theories.</p><p><strong>Lucas: </strong>So if I have 20 bad gradients and like, 10,000 good gradients for a video, why is it that with averaging 20 bad gradients are messing up the-</p><p><strong>El Mahdi: </strong>The amplitude. It’s like the billionaire in the room of core academics.</p><p><strong>Lucas: </strong>Okay, because the amplitude of each of their accounts is greater than the average of the other accounts?</p><p><strong>El Mahdi:</strong> Yes.</p><p><strong>Lucas:</strong> The average of other accounts that are going to engage with this thing don’t have as large of an amplitude because they haven’t engaged with this topic as much?</p><p><strong>El Mahdi: </strong>Yeah, because they’re not super credible on gun control, for example.</p><p><strong>Lucas: </strong>Yeah, but aren’t there a ton of other accounts with large amplitudes that are going to be looking at the same video and correcting over the-</p><p><strong>El Mahdi:</strong> Yeah, let’s define large amplitudes. If you come to the video and just like it, that’s a small update. What about you like it, post very engaging comments-</p><p><strong>Lucas:</strong> So you write a comment that gets a lot of engagement, gets a lot of likes and replies.</p><p><strong>El Mahdi: </strong>Yeah, that’s how you increase your amplitude. And because you are already doing some good job in becoming the reference on that video-sharing platform when it comes to discussing gun control, the amplitude of your commands is by definition high and the fact that your command was very early on posted and then not only you commented the video but you also produced a follow-up video.</p><p><strong>Lucas:</strong> I see, so the gradient is really determined by a multitude of things that the video-sharing platform is measuring for, and the metrics are like, how quickly you commented, how many people commented and replied to you. Does it also include language that you used?</p><p><strong>El Mahdi:</strong> Probably. It depends on the social media platform and it depends on the video-sharing platform and, what is clear is that there are many schemes that those 20 accounts created by this dozen people in a house can try to find good ways to maximize the amplitude of their generated gradients, but this is a way easier problem than the typical problems we have in technical AI safety. This is not value alignment or value loading or coherent extrapolated volition. This is a very easy, tractable problem on which now we have good news, provable results. What’s interesting is the follow-up questions that we are trying to investigate here with my colleagues, the first of which is, don’t necessarily have a majority of people on the internet promoting vaccines.</p><p><strong>Lucas:</strong> People that are against things are often louder than people that are not.</p><p><strong>El Mahdi:</strong> Yeah, makes sense, and sometimes maybe numerous because they generate content, and the people who think vaccines are safe not creating content. In some topics it might be safe to say that we have a majority of reasonable, decent people on the internet. But there are some topics in which now even like polls, like the vaccine situation, there’s a surge now of anti-vaccine resentment in western Europe and the US. Ironically this is happening in the developed country now, because people are so young, they don’t remember the non-vaccinated person. My aunt, I come from Morocco. my aunt is handicapped by polio, so I grew up seeing what a non-vaccinated person looks like. So young people in the more developed countries never had a living example of non-vaccinated past.</p><p><strong>Lucas: </strong>But they do have examples of people that end up with autism and it seems correlated with vaccines.</p><p><strong>El Mahdi: </strong>Yeah, the anti-vaccine content may just end up being so click baits, and so provocative that it gets popular. So this is a topic where the majority hypothesis which is crucial to poisoning resilience does not hold. An open follow up we’re onto now is how to combine ideas from reputation metrics, PageRank, et cetera, with poisoning resilience. So for example you have the National Health Institute, the John Hopkins Medical Hospital, Harvard Medical School, and I don’t know, the Massachusetts General Hospital having official accounts on some video-sharing platform and then you can spot what they say on some topic because now we are very good at doing semantic analysis of contents.</p><p>And know that okay, on the tag vaccines, I know that there’s this bunch of experts and then what you want to make emerge on your platform is some sort of like epistocracy. The power is given to the knowledgeable, like we have in some fields, like in medical regulation. The FDA doesn’t do a majority vote. We don’t have a popular majority vote across the country to tell the FDA whether it should approve this new drug or not. The FDA does some sort of epistocracy where the knowledgeable experts on the topic would vote. So how about mixing ideas from social choice?</p><p><strong>Lucas: </strong>And topics in which there are experts who can inform.</p><p><strong>El Mahdi: </strong>Yeah. There’s also a general fall-off of just straight out trying to connect Byzantine resilient learning with social choice, but then there’s another set of follow ups that motivates me even more. We were mentioning workers, workers, people generate accounts on social media, accounts generation gradients. That’s all I can implicitly assume in that the server, the abstraction that’s gathering those gradients is reliable. What about the aggregated platform itself being deployed on rogue machines? So imagine you are whatever platform doing learning. By the way, whatever always we have said from the beginning until now applies as long as you do gradient-based learning. So it can be recommended systems. It can be training some deep reinforcement learning of some super complicated tasks to beat, I don’t know the word, champion in poker.</p><p>We do not care as long as there’s some gradient generation from observing some state, some environmental state, and some reward or some label. It can be supervised, reinforced, as long as gradient based or what you say apply. Imagine now you have this platform leveraging distributed gradient creators, but then the platform itself for security reasons is deployed on several machines for fault tolerance. But then those machines themselves can fail. You have to make the servers agree on the model, so despite the fact that a fraction of the workers are not reliable and now a fraction of the servers themselves. This is the most important follow up i’m into now and I think there would be something on archive maybe in February or March on that.</p><p>And then a third follow up is practical instances of that, so I’ve been describing speculative thought experiments on power poisoning systems is actually brilliant master students working which means exactly doing that, like on typical recommended systems, datasets where you could see that it’s very easy. It really takes you a bunch of active agents to poison, a hundred thousand ones or more. Probably people working on big social media platforms would have ways to assess what I’ve said, and so as researchers in academia we could only speculate on what can go wrong on those platforms, so what we could do is just like we just took state of the art recommender systems, datasets, and models that are publicly available, and you can show that despite having a large number of reliable recommendation proposers, a small, tiny fraction of proposers can make, I don’t know, like a movie recommendation system recommend the most suicidal triggering film to the most depressed person watching through your platform. So I’m saying, that’s something you don’t want to have.</p><p><strong>Lucas: </strong>Right. Just wrapping this all up, how do you see this in the context of AI alignment and the future of machine learning and artificial intelligence?</p><p><strong>El Mahdi: </strong>So I’ve been discussing this here with people in the Beneficial AI conference and it seems that there are two schools of thought. I am still hesitating between the two because I switched within the past three months from the two sides like three times. So one of them thinks that an AGI is by definition resilient to poisoning.</p><p><strong>Lucas: </strong>Aligned AGI might be by definition.</p><p><strong>El Mahdi:</strong> Not even aligned. The second school of thought, aligned AGI is Byzantine resilient.</p><p><strong>Lucas:</strong> Okay, I see.</p><p><strong>El Mahdi:</strong> Obviously aligned AGI would be poisoning resilience, but let’s just talk about super intelligent AI, not necessarily aligned. So you have a super intelligence, would you include poisoning resilience in the super intelligence definition or not? And one would say that yeah, if you are better than human in whatever task, it means you are also better than human into spotting poison data.</p><p><strong>Lucas:</strong> Right, I mean the poison data is just messing with your epistemics, and so if you’re super intelligent your epistemics would be less subject to interference.</p><p><strong>El Mahdi:</strong> But then there is that second school of thought which I switched back again because I find that most people are in the first school of thought now. So I believe that super intelligence doesn’t necessarily include poisoning resilience because of what I call practically time constrained superintelligence. If you have a deadline because of computational complexity, you have to learn something, which can sometimes-</p><p><strong>Lucas: </strong>Yeah, you want to get things done.</p><p><strong>El Mahdi:</strong> Yeah, so you want to get it done in a finite amount of time. And because of that you will end up leveraging to speed up your learning. So if a malicious agent just put up bad observations of the environment or bad labeling of whatever is around you, then it can make you learn something else than what you would like as an aligned outcome. I’m strongly on the second side despite many disagreeing with me here. I don’t think super intelligence includes poisoning resilience, because super intelligence would still be built with time constraints.</p><p><strong>Lucas:</strong> Right. You’re making a tradeoff between safety and computational efficiency.</p><p><strong>El Mahdi: </strong>Right.</p><p><strong>Lucas: </strong>It also would obviously seem to matter the kind of world that the ASI finds itself in. If it knows that it’s in a world with no, or very, very, very few malevolent agents that are wanting to poison it, then it can just throw all of this out of the window, but the problem is that we live on a planet with a bunch of other primates that are trying to mess up our machine learning. So I guess just as a kind of fun example in taking it to an extreme, imagine it’s the year 300,000 AD and you have a super intelligence which has sort of spread across space-time and it’s beginning to optimize its cosmic endowment, but it gives some sort of uncertainty over space-time to whether or not there are other super intelligences there who might want to poison its interstellar communication in order to start taking over some of its cosmic endowment. Do you want to just sort of explore?</p><p><strong>El Mahdi:</strong> Yeah, that was like a closed experiment I proposed earlier to Carl Shulman from the FHI. Imagine some super intelligence reaching the planets where there is a smart form of life emerging from electric communication between plasma clouds. So completely non-carbon, non-silicon based.</p><p><strong>Lucas:</strong> So if Jupiter made brains on it.</p><p><strong>El Mahdi:</strong> Yeah, like Jupiter made brains on it just out of electric communication through gas clouds.</p><p><strong>Lucas: </strong>Yeah, okay.</p><p><strong>El Mahdi:</strong> And then this turned to a form of communication is smart enough to know that this is a super intelligence reaching the planet to learn about this form of life, and then it would just start trolling it.</p><p><strong>Lucas:</strong> It’ll start trolling the super intelligence?</p><p><strong>El Mahdi:</strong> Yeah. So they would come up with an agreement ahead of time, saying, “Yeah, this super intelligence coming from earth throughout our century to discover how we do things here. Let’s just behave dumbly, or let’s just misbehave. And then the super intelligence will start collecting data on this life form and then come back to earth saying, Yeah, they’re just a dumb plasma passive form of nothing interesting.</p><p><strong>Lucas:</strong> I mean, you don’t think that within the super intelligence’s model, I mean, we’re talking about it right now so obviously a super intelligence will know this when it leaves that there will be agents that are going to try and trick it.</p><p><strong>El Mahdi: </strong>That’s the rebuttal, yes. That’s the rebuttal again. Again, how much time does super intelligence have to do inference and draw conclusions? You will always have some time constraints.</p><p><strong>Lucas:</strong> And you don’t always have enough computational power to model other agents efficiently to know whether or not they’re lying, or …</p><p><strong>El Mahdi:</strong> You could always come up with thought experiment with some sort of other form of intelligence, like another super intelligence is trying to-</p><p><strong>Lucas:</strong> There’s never, ever a perfect computer science, never.</p><p><strong>El Mahdi:</strong> Yeah, you can say that.</p><p><strong>Lucas: </strong>Security is never perfect. Information exchange is never perfect. But you can improve it.</p><p><strong>El Mahdi: </strong>Yeah.</p><p><strong>Lucas: </strong>Wouldn’t you assume that the complexity of the attacks would also scale? We just have a ton of people working on defense, but if we have an equal amount of people working on attack, wouldn’t we have an equally complex method of poisoning that our current methods would just be overcome by?</p><p><strong>El Mahdi:</strong> That’s part of the empirical follow-up I mentioned. The one Isabella and I were working on, which is trying to do some sort of min-max game of poisoner versus poisoning resilience learner, adversarial poisoning setting where like a poisoner and then there is like a resilient learner and the poisoner tries to maximize. And what we have so far is very depressing. It turns out that it’s very easy to be a poisoner. Computationally it’s way easier to be the poisoner than to be-</p><p><strong>Lucas:</strong> Yeah, I mean, in general in the world it’s easier to destroy things than to create order.</p><p><strong>El Mahdi:</strong> As I said in the beginning, this is a sub-topic of technical AI safety where I believe it’s easier to have tractable formalizable problems for which you can probably have a safe solution.</p><p><strong>Lucas:</strong> Solution.</p><p><strong>El Mahdi:</strong> But in very concrete, very short term aspects of that. In March we are going to announce a major update in Tensor Flow which is the standout frameworks today to do distributed machine learning, open source by Google, so we will announce hopefully if everything goes right in sys ML in the systems for machine learning conference, like more empirically focused colleagues, so based on the algorithms I mentioned earlier which were presented at NuerIPS and ICML from the past two years, they will announce a major update where they basically changed every averaging insight in terms of flow by those three algorithms I mentioned, Krum and Bulyan and soon Kardam which constitute our portfolio of Byzantine resilience algorithms.</p><p>Another consequence that comes for free with that is that distributed machinery frameworks like terms of flow use TCPIP as a communication protocol. So TCPIP has a problem. It’s reliable but it’s very slow. You have to repeatedly repeat some messages, et cetera, to guarantee reliability, and we would like to have a faster communication protocol, like UDP. We don’t need to go through those details. But it has some package drop, so so far there was no version of terms of flow or any distributed machine learning framework to my knowledge using UDP. The old used TCPIP because they needed reliable communication, but now because we are Byzantine resilient, we can afford having fast but not completely reliable communication protocols like UDP. So one of the things that come for free with Byzantine resilience is that you can move from heavy-</p><p><strong>Lucas: </strong>A little bit more computation.</p><p><strong>El Mahdi: </strong>-yeah, heavy communication protocols like TCPIP to lighter, faster, more live communication protocols like UDP.</p><p><strong>Lucas: </strong>Keeping in mind you’re trading off.</p><p><strong>El Mahdi:</strong> Exactly. Now we have this portfolio of algorithms which can serve many other applications besides just making faster distributed machine learning, like making poisoning resilience. I don’t know, recommended systems for social media and hopefully making AGI learning poisoning resilience matter.</p><p><strong>Lucas: </strong>Wonderful. So if people want to check out some of your work or follow you on social media, what is the best place to keep up with you?</p><p><strong>El Mahdi: </strong>Twitter. My handle is El Badhio, so maybe you would have it written down on the description.</p><p><strong>Lucas:</strong> Yeah, cool.</p><p><strong>El Mahdi:</strong> Yeah, Twitter is the best way to get in touch.</p><p><strong>Lucas:</strong> All right. Well, wonderful. Thank you so much for speaking with me today and I’m excited to see what comes out of all this next.</p><p><strong>El Mahdi:</strong> Thank you. Thank you for hosting this.</p><p><strong>Lucas:</strong> If you enjoyed this podcast, please subscribe, give it a like, or share it on your preferred social media platform. We’ll be back again soon with another episode in the AI Alignment series.</p><p><em>[end of recorded material]</em></p><p><em>Originally published at </em><a href="https://futureoflife.org/2019/01/30/ai-alignment-podcast-the-byzantine-problem-poisoning-and-distributed-machine-learning-with-el-mahdi-el-mhamdi-beneficial-agi-2019/"><em>futureoflife.org</em></a><em> on January 30, 2019.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ccd71c18e388" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI Alignment Podcast: Cooperative Inverse Reinforcement Learning with Dylan Hadfield-Menell…]]></title>
            <link>https://medium.com/@FLIxrisk/ai-alignment-podcast-cooperative-inverse-reinforcement-learning-with-dylan-hadfield-menell-bbbf831740c0?source=rss-e33e2d2a809c------2</link>
            <guid isPermaLink="false">https://medium.com/p/bbbf831740c0</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ai-safety]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[ethics]]></category>
            <category><![CDATA[agi]]></category>
            <dc:creator><![CDATA[Future of Life]]></dc:creator>
            <pubDate>Thu, 17 Jan 2019 21:09:57 GMT</pubDate>
            <atom:updated>2019-01-21T16:06:01.212Z</atom:updated>
            <content:encoded><![CDATA[<h3>AI Alignment Podcast: Cooperative Inverse Reinforcement Learning with Dylan Hadfield-Menell (Beneficial AGI 2019)</h3><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fw.soundcloud.com%2Fplayer%2F%3Furl%3Dhttps%253A%252F%252Fapi.soundcloud.com%252Ftracks%252F560603265%26show_artwork%3Dtrue&amp;url=https%3A%2F%2Fsoundcloud.com%2Ffutureoflife%2Fcooperative-inverse-reinforcement-learning-with-dylan-hadfield-menell-beneficial-agi-2019&amp;image=http%3A%2F%2Fi1.sndcdn.com%2Fartworks-000473531712-z2i9dl-t500x500.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=soundcloud" width="800" height="166" frameborder="0" scrolling="no"><a href="https://medium.com/media/7874814695f379cc623d8714154be5b6/href">https://medium.com/media/7874814695f379cc623d8714154be5b6/href</a></iframe><p>What motivates cooperative inverse reinforcement learning? What can we gain from recontextualizing our safety efforts from the CIRL point of view? What possible role can pre-AGI systems play in amplifying normative processes?</p><p>Cooperative Inverse Reinforcement Learning with Dylan Hadfield-Menell is the eighth podcast in the AI Alignment Podcast series, hosted by Lucas Perry and was recorded at the Beneficial AGI 2018 conference in Puerto Rico. For those of you that are new, this series covers and explores the AI alignment problem across a large variety of domains, reflecting the fundamentally interdisciplinary nature of AI alignment. Broadly, Lucas will speak with technical and non-technical researchers across areas such as machine learning, governance, ethics, philosophy, and psychology as they pertain to the project of creating beneficial AI. If this sounds interesting to you, we hope that you will join in the conversations by following us or subscribing to our podcasts on <a href="https://www.youtube.com/channel/UC-rCCy3FQ-GItDimSR9lhzw">Youtube</a>, <a href="https://soundcloud.com/futureoflife/tracks">SoundCloud</a>, or your preferred podcast site/application.</p><p>If you’re interested in exploring the interdisciplinary nature of AI alignment, we suggest you take a look <a href="https://futureoflife.org/valuealignmentmap/">here at a preliminary landscape</a> which begins to map this space.</p><p>In this podcast, Lucas spoke with Dylan Hadfield-Menell. Dylan is a 5th year PhD student at UC Berkeley advised by Anca Dragan, Pieter Abbeel and Stuart Russell, where he focuses on technical AI alignment research.</p><p>Topics discussed in this episode include:</p><ul><li>How CIRL helps to clarify AI alignment and adjacent concepts</li><li>The philosophy of science behind safety theorizing</li><li>CIRL in the context of varying alignment methodologies and it’s role</li><li>If short-term AI can be used to amplify normative processes</li></ul><p><em>You can follow Dylan </em><a href="https://twitter.com/dhadfieldmenell"><em>here</em></a><em> and find the Cooperative Inverse Reinforcement Learning paper </em><a href="https://arxiv.org/abs/1606.03137"><em>here.</em></a><em> You can listen to the podcast above or read the transcript below.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6QWDrzeWSM_1wVVYRwhJZA.jpeg" /></figure><p><strong>Lucas:</strong> Hey everyone, welcome back to the AI Alignment Podcast series. I’m Lucas Perry and today we will be speaking for a second time with Dylan Hadfield-Menell on cooperative inverse reinforcement learning, the philosophy of science behind safety theorizing, CIRL in the context of varying alignment methodologies, and if short term AI can be used to amplify normative processes. This time it just so happened to be an in person discussion and Beneficial AGI 2019, FLI’s sequel to the Beneficial AI 2017 conference at Asilomar.</p><p>I have a bunch of more conversations that resulted from this conference to post soon and you can find more details about the conference in the coming weeks. As always, if you enjoy this podcast, please subscribe or follow us on your preferred listening platform. As many of you will already know, Dylan is a fifth year Ph.D. student at UC Berkeley, advised by Anca Dragan, Pieter Abbeel, and Stuart Russell, where he focuses on technical AI Alignment research. And so without further ado, I’ll give you Dylan.</p><p>Thanks so much for coming on the podcast again, Dylan, that’s been like a year or something. Good to see you again.</p><p><strong>Dylan:</strong> Thanks. It’s a pleasure to be here.</p><p><strong>Lucas:</strong> So just to start off, we can go ahead and begin speaking a little bit about your work on cooperative inverse reinforcement learning and whatever sorts of interesting updates or explanation you have there.</p><p><strong>Dylan: </strong>Thanks. For me, working in cooperative IRL has been a pretty long process, it really sort of dates back to the start of my second year in PhD when my advisor came back from a yearlong sabbatical and suggested that we entirely changed the research direction we were thinking about.</p><p>That was to think about AI Alignment and AI Safety and associated concerns that, that might bring. And our first attempt at a really doing research in that area was to try to formalize what’s the problem that we’re looking at, what are the space of parameters and the space of solutions that we should be thinking about in studying that problem?</p><p>And so it led us to write Cooperative Inverse Reinforcement Learning. Since then I’ve had a large amount of conversations where I’ve had incredible difficulty trying to convey what it is that we’re actually trying to do here and what exactly that paper and idea represents with respect to AI Safety.</p><p>One of the big updates for me and one of the big changes since we’ve spoken last, is getting a little bit of a handle on really what’s the value of that as the system. So for me, I’ve come around to the point of view that really what we were trying to do with cooperative IRL was to propose an alternative definition of what it means for an AI system to be effective or rational in some sense.</p><p>And so there’s a story you can tell about artificial intelligence, which is that we started off and we observed that people were smart and they were intelligent in some way, and then we observed that we could get computers to do interesting things. And this posed the question of can we get computers to be intelligent? We had no idea what that meant, no idea how to actually nail it down and we discovered that in actually trying to program solutions that looked intelligent, we had a lot of challenges.</p><p>So one of the big things that we did as a field was to look over next door into the economics department in some sense, to look at those sort of models that they have of decision theoretic rationality and really looking at homoeconomicous as an ideal to shoot for. From that perspective, actually a lot of the field of AI has shifted to be about effective implementations of homoeconomicous.</p><p>In my terminology, this is about systems that are effectively individually rational. These are systems that are good at optimizing for their goals, and a lot of the concerns that we have about AI Safety is that systems optimizing for their own goals could actually lead to very bad outcomes for the rest of us. And so what cooperative IRL attempts to do is to understand what it would mean for a human robot system to behave as a rational agent.</p><p>In the sense, we’re moving away from having a box drawn around the AI system or the artificial component of the system to having that agent box drawn around the person and the system together, and we’re trying to model the sort of important parts of the value alignment problem in our formulation here. And in this case, we went with the simplest possible set of assumptions which are basically that we have a static set of preferences that are the humans preferences that they’re trying to optimize. This is effectively the humans welfare.</p><p>The world is fully observable and the robot and the person are both working to maximize the humans welfare, but there is this information bottlenecking. This information asymmetry that’s present that we think is a fundamental component of the value alignment problem. And so really what cooperative IRL, is it’s a definition of how a human and a robot system together can be rational in the context of fixed preferences in a fully observable world state.</p><p><strong>Lucas:</strong> There’s a point of metatheory or coming up with models and theory. It seems like the fundamental issue is given how and just insanely complex AI Alignment is trying to converge on whatever the most efficacious model is, is very, very difficult. People keep flicking back and forth about theoretically how we’re actually going to do this. Even in very grid world or toy environments. So it seemed very, very hard to isolate the best variables or what variables can be sort of modeled and tracked in ways that is going to help us most.</p><p><strong>Dylan:</strong> So, I definitely think that this is not an accurate model of the world and I think that there are assumptions here which, if not appropriately reexamined, would lead to a mismatch between the real world and things that work in theory.</p><p><strong>Lucas:</strong> Like human beings having static preferences.</p><p><strong>Dylan:</strong> So for example, yes, I don’t claim to know what human preferences really are and this theory is not an attempt to say that they are static. It is an attempt to identify a related problem to the one that we’re really faced with, that we can actually make technical and theoretical progress on. That will hopefully lead to insights that may transfer out towards other situations.</p><p>I certainly recognize that what I’m calling a theta in that paper is not really the same thing that everyone talks about when we talk about preferences. I, in talking with philosophers, I’ve discovered, I think it’s a little bit more closer to things like welfare in like a moral philosophy context, which maybe you could think about as being a more static object that you would want to optimize.</p><p>In some sense theta really is an encoding of what you would like the system to do, in general is what we’re assuming there.</p><p><strong>Lucas: </strong>Because it’s static.</p><p><strong>Dylan:</strong> Yes, and to the extent that you want to have that be changing over time, I think that there’s an interesting theoretical question as to how that actually is different, and what types of changes that leads to and whether or not you can always reduce something with non-static preferences to something with static preferences from a mathematical point of view.</p><p><strong>Lucas:</strong> I can see how moving from static to changing over time just makes it so much more insanely complex.</p><p><strong>Dylan: </strong>Yeah, and it’s also really complex of the level of its Philosophically unclear what the right thing to do.</p><p><strong>Lucas:</strong> Yeah, that’s what I mean. Yeah, you don’t even know what it even means to be aligning as the values are changing, like whether or not the agent even thinks that they just moved in the right direction or not.</p><p><strong>Dylan: </strong>Right, and I also even think I want to point out how uncertain all of these things are. We as people are hierarchical organizations have different behaviors and observation systems and perception systems. And we believe we have preferences, we have a name to that, but there is a sense in which that is ultimately a fiction of some kind.</p><p>It’s a useful tool that we have to talk about ourselves to talk about others that facilitates interaction and cooperation. And so given that I do not know the answer to these philosophical questions, what can I try to do as a technical researcher to push the problem forward and to make actual progress?</p><p><strong>Lucas: </strong>Right, and so it’s sort of again, like a metatheoretical point and what people are trying to do right now in the context of AI Alignment, it seems that the best thing for people to be doing is sort of to be coming up with these theoretical models and frameworks, which have a minimum set of assumptions which may be almost like the real world but are not, and then making theoretical progress there that will hopefully in the future transfer, as you said to other problems as ML and deep learning gets better and the other tools are getting better so that it’ll actually have the other tools to make it work with more complicated assumptions.</p><p><strong>Dylan:</strong> Yes, I think that’s right. The way that I view this as we had AI, is this broad, vague thing. Through the course of AI research, we kind of got to Markov decision processes as a sort of coordinating theory around what it means for us to design good agents, and cooperative IRL is an attempt to take a step from markup decision processes more closely towards the set of problems that we want to study.</p><p><strong>Lucas:</strong> Right, and so I think this is like a really interesting point that I actually haven’t talked to anyone else about and if you have a few more words about it, I think it would be really interesting. So just in terms of being a computer scientist and being someone who is working on the emerging theory of a field. I think it’s often unclear what the actual theorizing process is behind how people get to CIRL. How did someone get to debate? How did someone get to iterated amplification?</p><p>It seems like you first identify problems which you see to be crucial and then there are some sorts of epistemic and pragmatic heuristics that you apply to try and begin to sculpt a model that might lead to useful insight. Would you have anything to correct or unpack here?</p><p><strong>Dylan: </strong>I mean, I think that is a pretty good description of a pretty fuzzy process.</p><p><strong>Lucas: </strong>But like being a scientist or whatever?</p><p><strong>Dylan: </strong>Yeah. I don’t feel comfortable speaking for scientists in general here, but I could maybe say a little bit more about my particular process, which is that I try to think about how I’m looking at the problem differently from other people based on different motivations and different goals that I have. And I try to lean into how that can push us in different directions. There’s a lot of other really, really smart people who have tried to do lots of things.</p><p>You have to maintain an amount of intellectual humility about your ability to out think the historical components of the field. And for me, I think that in particular for AI Safety, it’s thinking about reframing what is the goal that we’re shooting towards as a field.</p><p><strong>Lucas: </strong>Which we don’t know.</p><p><strong>Dylan: </strong>We don’t know of those goals are, absolutely. And I think that there is a sense in which the field has not re-examined those goals incredibly deeply. For a little bit, I think that it’s so hard to do anything that looks intelligent in the real world that we’ve been trying to focus on that individually rational Markov decision process model. And I think that a lot of the concerns about AI Safety are really a call for AI as a field to step back and think about what we’re trying to accomplish in the world and how can we actually try to achieve beneficial outcomes for society.</p><p><strong>Lucas: </strong>Yeah, and I guess like a sociological phenomenon within the scientists or people who are committed to empirical things. In terms of reanalyzing what the goal of AI Alignment is, the sort of area of moral philosophy and ethics and other things, which for empirical leaning rational people can be distasteful because you can’t just take a telescope to the universe and see like a list of what you ought to do.</p><p>And so it seems like people like to defer on these questions. I don’t know. Do you have anything else to add here?</p><p><strong>Dylan:</strong> Yeah. I think computer scientists in particular are selected to be people who like having boxed off problems that they know how to solve and feel comfortable with, and that leaning into getting more people with a humanities bent into computer science and broadly AI in particular, AI Safety especially is really important and I think that’s a broad call that we’re seeing come from society generally.</p><p><strong>Lucas: </strong>Yeah, and I think it also might be wrong though to model the humanities questions as those which are not in boxes and cannot be solved. That’s sort of like a logical positivist thing to say, that on one end we have the hard things and you just have to look at the world enough and you’ll figure it out and then there’s the soft squishy things which deal with abstractions that I don’t have real answers, but people with fluffy degrees need to come up with things that seem right but aren’t really right.</p><p><strong>Dylan: </strong>I think it would be wrong to take what I just said in that direction, and if that’s what it sounds like I definitely want to correct that. I don’t think there is a sense in which computer science is a place where there are easy right answers, and that the people in humanities are sort of waving their hands and sort of fluffing around.</p><p>This is sort of leaning into making this a more AI value alignment kinds of framing or thinking about it. But when I think about being AI systems into the world, I think about what things can you afford to get wrong in your specification and which things can you not afford to get wrong in your specifications.</p><p>In this sense, specifying physics incorrectly is much, much better than specifying the objective incorrectly, at least by default. And the reason for that is what happens to the world when you push it, is a question that you can answer from your observations. And so if you start off in the wrong place, as long as you’re learning and adapting, I can reasonably expect my systems do correct to that. Or at least the goal of successful AI research is that your systems will effectively adapt to that.</p><p>However, the past that your system is supposed to do is sort of arbitrary in a very fundamental sense. And from that standpoint, it is on you as the system designer to make sure that objective is specified correctly. When I think about what we want to do as a field, I ended up taking a similar lens and that there’s a sense in which we as researchers and people and society and philosophers and all of it are trying to figure out what we’re trying to do and what we want to task the technology with, and the directions that we want to push it in. And then there are questions of what will the technology be like and how should it function that will be informed by that and shaped by that.</p><p>And I think that there is a sense in which that is arbitrary. Now, what is right? That I don’t really know the answer to and I’m interested in having those conversations, but they make me feel uneasy. I don’t trust myself on those questions, and that could mean that I should learn how to feel more uneasy and think about it more and in doing this research I have been kind of forced into some of those conversations.</p><p>But I also do think that for me at least I see a difference between what can we do and what should we do. And thinking about what should we do as a really, really hard question that’s different than what can we do.</p><p><strong>Lucas: </strong>Right. And so I wanna move back towards CIRL, but just to sort of wrap up here on our philosophy of science musings, a thought I had while you were going through that was, at least for now, what I think is fundamentally shared between fields that deal with things that matter, are their concepts deal with meaningfully relevant reference in the world? Like do your concepts refer to meaningful things?</p><p>Putting ontology aside, whatever love means or whatever value alignment mean. These are meaningful referents for people and I guess for now if our concepts are actually referring to meaningful things in the world, then it seems important.</p><p><strong>Dylan: </strong>Yes, I think so. Although, I’m not totally sure I understood that.</p><p><strong>Lucas: </strong>Sure, that’s fine. People will say that humanities or philosophy doesn’t have these boxes with like well-defined problems and solutions because they either don’t deal with real things in the world or the concepts are so fuzzy that the problems are sort of invented and illusory. Like how many angels can stand on the head of a pin? Like the concepts don’t work, aren’t real and don’t have real referents, but whatever.</p><p>And I’m saying the place where philosophy and ethics and computer science and AI Alignment should at least come together for now is where the referents have, where the concepts of meaningful referents in the world?</p><p><strong>Dylan: </strong>Yes, that is something that I absolutely buy. Yes, I think there’s a very real sense in which those questions are harder, but that doesn’t mean they’re less real or less important.</p><p><strong>Lucas:</strong> Yes, that’s because it’s the only point I wanted to push against logical positivism.</p><p><strong>Dylan:</strong> No, I don’t mean to say that the answers are wrong, it’s just that they are harder to prove in a real sense.</p><p><strong>Lucas: </strong>Yeah. I mean, I don’t even know if they have answers or if they do or if they’re just all wrong, but I’m just open to it and like more excited about everyone coming together thing.</p><p><strong>Dylan: </strong>Yes, I absolutely agree with that.</p><p><strong>Lucas: </strong>Cool. So now let’s turn it back into the CIRL. So you began by talking about how you and your advisers were having this conceptual shift and framing, then we got into the sort of philosophy of science behind how different models and theories of alignment go. So from here, whatever else you have to say about CIRL.</p><p><strong>Dylan: </strong>So I think for me the upshot of concerns about advanced AI systems and negative consequence there in really is a call to recognize that the goal of our field is AI Alignment. That almost any AI that’s not AI Alignment is solving a sub problem and viewing it only in solving that sub problem is a mistake.</p><p>Ultimately, we are in the business of building AI systems that integrate well with humans and human society. And if we don’t take that as a fundamental tenant of the field, I think that we are potentially in trouble and I think that that is a perspective that I wish was more pervasive throughout artificial intelligence generally,</p><p><strong>Lucas:</strong> Right, so I think I do want to move into this view where safety is a normal thing, and like Stuart Russell will say, “People who build bridges all care about safety and there aren’t a subsection of bridge builders who work in bridge safety, everyone is part of the bridge safety.” And I definitely want to get into that, but I also sort of want to get a little bit more into CIRL and why you think it’s so motivating and why this theoretical framing and shift is important or illuminating, and what the specific content of it is.</p><p><strong>Dylan:</strong> The key thing is that what it does is point out that it doesn’t make sense to talk about how well your system is doing without talking about the way in which it was instructed and the type of information that it got. No AI system exists on its own, every AI system has a designer, and it doesn’t make sense to talk about the functioning of that system without also talking about how that designer built it, evaluated it and how well it is actually serving those ends.</p><p>And I don’t think this is some brand new idea that no one’s ever known about, I think this is something that is incredibly obvious to practitioners in the field once you pointed out. The process whereby a robot learns to navigate a maze or vacuum a room is not, there is an objective and it optimizes it and then it does it.</p><p>What it is that there is a system designer who writes down an objective, selects an optimization algorithm, observes the final behavior of that optimization algorithm, goes back, modifies the objectives, modifies the algorithm, changes hyper parameters, and then runs it again. And there’s this iterative process whereby your system eventually ends up getting to the behavior that you wanted to have. And AI researchers have tended to draw a box around. The thing that we call AI is the sort of final component of that.</p><p><strong>Lucas: </strong>Yeah, it’s because at least subjectively and I guess this is sort of illuminated by meditation and Buddhism, is that if you’re a computer scientist and you’re just completely identified with the process of doing computer science, you’re just identified with the problem. And if you just have a little bit of mindfulness and you’re like, “Okay, I’m in the context of a process where I’m an agent and trying to align another agent,” and if you’re not just completely identified with the process and you see the unfolding of the process, then you can do sort of like more of a meta-analysis which takes a broader view of the problem and can then, I guess hopefully work on improving it.</p><p><strong>Dylan:</strong> Yeah, I think that’s exactly right, or at least as I understand that, that’s exactly right. And to be a little bit specific about this, we have had these engineering principles and skills that are not in the papers, but they are things that are passed down from Grad student to Grad student within a lab. Their institutional knowledge that exists within a company for how you actually verify and validate your systems, and cooperative IRLs and attempt to take all of that sort of structure that AI systems have existed within and try to bring that into the theoretical frameworks that we actually work with.</p><p><strong>Lucas:</strong> So can you paint a little picture of what the CIRL model looks like?</p><p><strong>Dylan:</strong> It exists in a sequential decision making context and we assume we have states of the world and a transition diagram that basically tells us how we get to another state given the previous state and actions from the human and the robot. But the important conceptual shift that it makes is the space of solutions that we’re dealing with are combinations of a teaching strategy and a learning strategy.</p><p>There is a commitment on the side of the human designers or users of the systems to provide data that is in some way connected to the objectives that they want to be fulfilled. That data can take many forms, it could be in the form of writing down a reward function that ranks a set of alternatives, it could be in the form of providing demonstrations that you expect your system to imitate. It could be in the form of providing binary comparisons between two clearly identified alternatives.</p><p>And the other side of the problem is what is the learning strategy that we use? And this is the question of how the robot is actually committing to respond to the observations that we’re giving it about what we wanted to do, in the case of a pre-specified proxy reward going to a literal interpretation and a reinforcement learning system, let’s say. What the system is committing to doing is optimizing under that set of trajectory rankings and preferences based off the simulation environment that it’s in, or the actual physical environment that it’s exploring.</p><p>When we shift to something like inverse reward design, which is a paper that we released last year, what that says is we’d like the system to look at this ranking of alternatives and actually try to blow that up into a larger uncertainty set over the set of possible consistent rankings with that, and then when you go into deployment, you may be able to leverage that uncertainty to avoid catastrophic failures or generally just unexpected behavior.</p><p><strong>Lucas: </strong>So this other point I think that you and I discussed briefly, maybe it was actually with Rohan, but it seems like often in terms of AI Alignment, it’s almost like we’re reasoning from nowhere about abstract agents and that sort of makes the problem extremely difficult. Often, if you just look at human examples, it just becomes super mundane and easy. This sort of conceptual shift can almost I think be framed super simply as like the difference between a teacher trying to teach someone and then a teacher realizing that the teacher is a person that is teaching another student and the teacher can think better about how to teach and then also the process between the teacher and the student and how to improve that at a higher level of attraction.</p><p><strong>Dylan:</strong> I think that’s the direction that we’re moving in. What I would say is it’s as AI practitioners, we are teaching our systems how to behave and we have developed our strategies for doing that.</p><p>And now that we’ve developed a bunch of strategies that sort of seem to work. I think it’s time for us to develop a more rigorous theory of actually how those teaching strategies interact with the final performance of the system.</p><p><strong>Lucas: </strong>Cool. Is there anything else here that you would like say about CIRL, or any really important points you would like people to get people who are interested in technical AI Alignment or CS students?</p><p><strong>Dylan: </strong>I think the main point that I would make is that research and thinking about powerful AI systems is valuable, even if you don’t think that that’s what’s going to happen. You don’t need to be motivated by those sets of problems in order to recognize that this is actually just basic research into the science of artificial intelligence.</p><p>It’s got an incredible amount of really interesting problems and the perspectives that you adopt from this framing can be incredibly useful as a comparative advantage over other researchers in the field. I think that’d be my final word here.</p><p><strong>Lucas: </strong>If I might just ask you one last question. We’re at beneficial AGI 2019 right now and we’ve heard a lot of overviews of different research agendas and methodologies and models and framings for how to best go forth with AI Alignment, which include a vast range of things which work on corrigibility and interpretability and robustness and other things, and the different sort of research agendas and methodologies of places like MIRI who is come out with this new framing on embedded agency, and also different views at OpenAI and DeepMind.</p><p>And Eric Drexler has also newly proposed these services based conception of AI where we remove the understanding of powerful AI systems or regular AI systems as agents, which sort of gets us away from a lot of the x-risky problems and global catastrophic risks problems and value alignment problems.</p><p>From your point of view, as someone who’s worked a lot in CIRL and is the technical alignment researcher, how do you view CIRL in this context and how do you view all of these different emerging approaches right now in AI Alignment?</p><p><strong>Dylan: </strong>For me, and you know, I should give a disclaimer. This is my research area and so I’m obviously pretty biased to thinking it’s incredibly important and good, but for me at least, cooperative IRL is a uniting framework under which I can understand all of those different approaches. I believe that a services type solution to AI Safety or AI Alignment that’s actually arguing for a particular type of learning strategy and implementation strategies of CIRL, and I think it can be framed within that system.</p><p>Similarly, I had some conversations with people about debate. I believe debate fits really nicely into the framework and we commit to a human strategy of judging debates from systems and we commit to a robot strategy and just putting yourself into two systems and working towards that direction. So for me, it’s a way in which I can sort of identify the commonalities between these different approaches and compare and contrast them and then under a set of assumptions about what the world is like, what the space of possible preferences is like and what the space of strategies that people can implement possibly get out some information about which one is better or worse, or which type of strategy is vulnerable to different types of mistakes or errors.</p><p><strong>Lucas: </strong>Right, so I agree with all of that, the only place that I might want to push back is, it seems that maybe the MIRI embedded agency stuff subsumes everything else. What do you think about that?</p><p>Because the framing is like whenever AI researchers draw these models, there are these conceptions of these information channels, right, which are selected by the researchers and which we control, but the universe is really just a big non-dual happening of stuff and agents are embedded in the environment and are almost an identical process within the environment and it’s much more fuzzy where the dense causal streams are and where a little causal streams are and stuff like that. It just seems like the MIRI stuff seems to maybe subsume the CIRL and everything else a little bit more, but I don’t know.</p><p><strong>Dylan: </strong>I certainly agree that that’s the one that’s hardest to fit into the framework, but I would also say that in my mind, I don’t know what an agent is. I don’t know how to operationalize an agent, I don’t actually know what that means in the physical world and I don’t know what it means to be an agent. What I do know is that there is a strategy of some sort that we can think of as governing the ways that the system is perform and behave.</p><p>I want to be very careful about baking in assumptions in beforehand. And it feels to me like embedded agency is something that I don’t fully understand the set of assumptions being made in that framework. I don’t necessarily understand how they relate to the systems that we’re actually going to build.</p><p><strong>Lucas: </strong>When people say that an agent is like a fuzzy concept, I think that, that might be surprising to a lot of people who have thought somewhat about the problem because it’s like, obviously I know what an agent is, it’s different than all the other dead stuff in the world that has goals and it’s physically confined and unitary.</p><p>If you just like imagine like abiogenesis, how life began. It is the first relatively self-replicating chain of hydrocarbons and agent and you can go from a really small systems to really big systems, which can exhibit certain properties or principles that feel a little bit agenty, but may not be useful. And so I guess if we’re going to come up with a definition of it, it should just be something useful for us or something.</p><p><strong>Dylan: </strong>I think I’m not sure is the most accurate word we can use here. I wish I had a better answer for what this was, maybe I can share one of the thought experiments that convinced me, I was pretty confused about what an agent is.</p><p><strong>Lucas: </strong>Yeah, sure.</p><p><strong>Dylan: </strong>It came from thinking about what value alignment is. So if we think about values alignment between two agents and those are both perfectly rational actors, making decisions in the world perfectly in accordance with their values, with full information. I can sort of write down a definition of value alignment, which is basically you’re using the same ranking over alternatives that I am.</p><p>But a question that we really wanted to try to answer that feels really important is what does it mean to be value aligned in a partial context? If you were a bounded agent, if you’re not a perfectly rational agent, what does it actually mean for you to be value aligned? That was the question that we also didn’t really know how to answer.</p><p><strong>Lucas: </strong>My initial reaction is the kind of agent that tries its best with its limited rationality to be like the former thing that you talked about.</p><p><strong>Dylan: </strong>Right, so that leads to a question that we thought about, so as opposed I have a chess playing agent and it is my chess playing agent and so I wanted to win the game for me. Suppose it’s using the correct goal test, so it is actually optimizing for my values. Let’s say it’s only searching out to depth three, so it’s pretty dumb as far as chess players go.</p><p>Do I think that that is an agent that is value aligned with me? Maybe. I mean, certainly I can tell the story in one way that it sounds like it is. It’s using the correct objective function, it’s doing some sort of optimization thing. If it ever identifies a checkmate move in three moves, I will always find that get that back to me. And so that’s a sense in which it feels like it is a value aligned agent.</p><p>On the other hand, what if it’s using a heuristic function which is chosen poorly, or and something closer to an adversarial manner. So now it’s a depth three agent that is still using the correct goal test, but it’s searching in a way that is adversarially selected. Is that a partially value aligned agent?</p><p><strong>Lucas: </strong>Sorry, I don’t understand what it means to have the same objective function, but be searching in three depth in an adversarial way.</p><p><strong>Dylan:</strong> In particular, when you’re doing a chess search engine, there is your sort of goal tests that you run on your leaves of your search to see if you’ve actually achieved winning the game. But because you’re only doing a partial search, you often have to rely on using a heuristic of some sort to like rank different positions.</p><p><strong>Lucas: </strong>To cut off parts of the tree.</p><p><strong>Dylan: </strong>Somewhat to cut off parts of the tree, but also just like you’ve got different positions, neither of which are winning and you need to choose between those.</p><p><strong>Lucas: </strong>All right. So there’s a heuristic, like it’s usually good to take the center or like the queen is something that you should always probably keep.</p><p><strong>Dylan: </strong>Or these things that are like values of pieces that you can add up was I think one of the problems …</p><p><strong>Lucas: </strong>Yeah, and just as like an important note now in terms of the state of machine learning, the heuristics are usually chosen by the programmer. Are system is able to collapse on heuristics themselves?</p><p><strong>Dylan: </strong>Well, so I’d say one of the big things in like AlphaZero or AlphaGo as an approach is that they applied sort of learning on the heuristic itself and they figured out a way to use the search process to gradually improve the heuristic and have the heuristic actually improving the search process.</p><p>And so there’s sort of a feedback loop set up in those types of expert iteration systems. What my point here is that when I described that search algorithm to you, I didn’t mention what heuristic it was using at all. And so you had no reason to tell me whether or not that system was partially value aligned or not because actually with heuristic is 100 percent of what’s going to determine the final performance of the system and whether or not it’s actually helping you.</p><p>And then the sort of final point I have here that I might be able to confuse you with a little bit more is, what if we just sort of said, “Okay, forget this whole searching business. I’m just going to precompute all the solutions from my search algorithm and I’m going to give you a policy of when you’re in this position, do this move. When you’re in that position, do that move.” And what would it mean for that policy to be values aligned with me?</p><p><strong>Lucas: </strong>If it did everything that you would have done if you were the one playing the chess game. Like is that value alignment?</p><p><strong>Dylan: </strong>That certainly perfect imitation, and maybe we [crosstalk 00:33:04]</p><p><strong>Lucas: </strong>Perfect imitation isn’t necessarily value alignment because you don’t want it to perfectly imitate you, you want it to win the game.</p><p><strong>Dylan: </strong>Right.</p><p><strong>Lucas:</strong> Isn’t the easiest way to just sort of understand this is that there are degrees of value alignment and value alignment is the extent to which the thing is able to achieve the goals that you want?</p><p><strong>Dylan:</strong> Somewhat, but the important thing here is trying to understand what these intuitive notions that we’re talking about actually mean for the mathematics of sequential decision making. And so there’s a sense in which you and I can talk about partial value alignment and the agents that are trying to help you. But if we actually look at the math of the problem, it’s actually very hard to understand how that actually translates. Like mathematically I have lots of properties that I could write down and I don’t know which one of those I want to call partial value alignment.</p><p><strong>Lucas: </strong>You know more about the math than I do, but the percentage chance of a thing achieving the goal is the degree to which its value aligned? If you’re certain that the end towards which is striving, and the end towards what you want it to strive?</p><p><strong>Dylan:</strong> Right, but that striving term is a hard one, right? Because if your goals aren’t achievable then it’s impossible to be value aligned with you in that sense.</p><p><strong>Lucas: </strong>Yeah, you have to measure the degree to which the end towards which it’s striving is the end towards what you want it to strive and then also measure the degree to which the way that it tries to get to what you want is efficacious or …</p><p><strong>Dylan: </strong>Right. I think that intuitively I agree with you and I know what you mean, but it’s like I can do things like I can write down a reward function and I can say how well does this system optimize that reward function? And we could ask whether or not that means its value aligned with it or not. But to me, that just sounds like the question of like is your policy optimal and the sort of more standard context.</p><p><strong>Lucas: </strong>All right, so have you written about how you think that CIRL subsumes all of these other methodologies? And if it does subsume these other AI Alignment methodologies. How do you think that will influence or affect the way we should think about the other ones?</p><p><strong>Dylan: </strong>I haven’t written that explicitly, but when I’ve tried to convey is that it’s a formalization of the type of problem we’re trying to solve. I think describing this subsuming them is not quite right.</p><p><strong>Lucas: </strong>It contextualizes them and it brings light to them by providing framing.</p><p><strong>Dylan: </strong>It gives me a way to compare those different approaches and understand what’s different and what’s the same between them, and in what ways are they … like in what scenarios do we expect them to work out versus not? One thing that we’ve been thinking about recently is what happens when the person doesn’t know immediately and what they’re trying to do.</p><p>So if we imagine that there is in fact the static set of preferences, the person’s trying to optimize, so we’re still making that assumption, but assuming that those preferences are revealed to the person over time through experience or interaction with the world. That is a richer class of value alignment problems than cooperative IRL deals with. It’s really closer to what we are attempting to do right now.</p><p><strong>Lucas: </strong>Yeah, and I mean that doesn’t even include value degeneracy, like what if I get hooked on drugs in the next three years and all my values go and my IRL agent works on assumptions that I’m always updating towards what I want, but you know …</p><p><strong>Dylan: </strong>Yes, and I think that’s where you get these questions of changing preferences that make it hard to really think through things. I think there’s a philosophical stance you’re taking there, which is that your values have changed rather than your beliefs have changed there.</p><p>In the sense that wire-heading is a phenomenon that we see in people and in general learning agents, and if you are attempting to help it learning agent, you must be aware of the fact that wire-heading is a possibility and possibly bad. And then it’s incredibly hard to distinguish from someone who’s just found something that they really like and want to do.</p><p>When you should make that distinction or how you should make that distinction is a really challenging question, that’s not a purely technical computer science question.</p><p><strong>Lucas: </strong>Yeah, but even at the same time, I would like to demystify it a bit. If your friend got hooked on drugs, it’s pretty obvious for you why it’s bad, it’s bad because he’s losing control, it’s bad because he’s sacrificing all of his other values. It’s bad because he’s shortening his life span by a lot.</p><p>I just mean to win again, in this way, it’s obvious in ways in which humans do this, so I guess if we take biological inspired approaches to understanding cognition and transferring how humans deal with these things into AI machines, at least at face value seems like a good way of doing it, I guess.</p><p><strong>Dylan: </strong>Yes, everything that you said I agree with. My point is that those are in a very real sense, normative assumptions that you as that person’s friend are able to bring to the analysis of that problem, and in in some ways there is an arbitrariness to labeling that as bad.</p><p><strong>Lucas: </strong>Yeah, so the normative issue is obviously very contentious and needs to be addressed more, but at the same time society has come to very clear solutions to normative problems like murder is basically a solved normative problem. There’s a degree to which it’s super obvious that certain normative questions are just answer it and we should I guess practice epistemic humility and whatever here obviously.</p><p><strong>Dylan: </strong>Right, and I don’t disagree with you on that point, but I think what I’d say is, as a research problem there’s a real question to getting a better understanding of the normative processes whereby we got to solving that question. Like what is the human normative process? It’s a collective societal system. How does that system evolve and change? And then how should machines or other intelligent entities integrate into that system without either subsuming or destroying it in bad ways? I think that’s what I’m trying to get at when I make these points. There is something about what we’re doing here as a society that gets us to labeling these things in the ways that we do and calling them good or bad.</p><p>And on the one hand, as a person believe that there are correct answers and I know what I think is right versus what I think is wrong. And then as a scientist I want to try to take a little bit more of an outside view and try to understand like what is the process whereby we as a society or as genetic beings started doing that? Understanding what that process is and how that process evolves, and actually what that looks like in people now is a really critical research program.</p><p><strong>Lucas:</strong> So one thing that I tried to cover in my panel yesterday on what civilization should strive for, is in the short, medium, to longterm the potential role that narrow to general AI systems might play in amplifying human moral decision making.</p><p>Solving as you were discussing this sort of deliberative, normative process that human beings undergo to total converge on an idea. I’m just curious to know like with more narrow systems, if you’re optimistic about ways in which AI can sort of help and elucidate our moral decision making at work to amplify it.</p><p>And before I let you start, I guess there’s one other thing that I said that I think Rohin Shah pointed out to me that was particularly helpful in one place. But beyond the moral decision making, the narrow AI systems can help us by making the moral decision make, the decisions that we implement them faster than we could.</p><p>Depending on the way a self-driving car decides to crash is like an expression of our moral decision making in like a fast computery way. I’m just saying like beyond ways in which AI systems make moral decisions for us faster than we can, I don’t know, maybe in courts or other things which seem morally contentious. Are there also other ways in which they can actually help the deliberative process examining massive amounts of moral information or like a value information or analyzing something like an aggregated well-being index where we try to understand more so how policies impact the wellbeing of people or like what sorts of moral decisions lead to good outcomes, whatever. So whatever you have to say to that.</p><p><strong>Dylan: </strong>Yes, I definitely want to echo that. We can sort of get a lot of pre-deliberation into a fast timescale reaction with AI systems and I think that that is a way for us to improve how we act in the quality of the things that we do from a moral perspective. That you do see a real path and to actually bringing that to be in the world.</p><p>In terms of helping us actually deliberate better, I think that is a harder problem that I think is absolutely worth more people thinking about but I don’t know the answers here. What I do think is that if we have a better understanding of what the deliberative process is, I think there are correct questions to look at to try to get to that or not, the moral questions about what’s right and what’s wrong and what do we think is right and what do we think is wrong, but they are much more questions at the level of what is it about our evolutionary pathway that led us to thinking that these things are right or wrong.</p><p>What is it about society and the pressures that you’re gone and faced that led us to things where murder is wrong in almost every society in the world. I will say the death penalty is the thing, it’s just the type of sanctioned murder. So there is a sense in which I think it’s a bit more nuanced than just that. And there’s something to be said about like I guess if I had to make my claims, like what I think has sort of happened there.</p><p>So there’s something about us as creatures that evolved to coordinate and perform well in groups and pressures that, that placed on us that caused us to develop these normative systems whereby we say different things are right and wrong.</p><p><strong>Lucas: </strong>Iterated game theory over millions of years or something.</p><p><strong>Dylan: </strong>Something like that. Yeah, but there’s a sense in which us labeling things as right and wrong and developing the processes whereby we label things as right and wrong is a thing that we’ve been pushed towards.</p><p><strong>Lucas:</strong> From my perspective, it feels like this is more tractable than people lead on, like AI is only going to be able to help in moral deliberation, once it’s general. It already helps us in regular deliberation and moral deliberation isn’t a special kind of deliberation and moral deliberation requires empirical facts about the world and in persons just like any other kind of actionable deliberation does and domains that aren’t considered to have to do with moral philosophy or ethics or things like that.</p><p>So I’m not an AI researcher, but it seems to me like this is more attractable than people lead onto be. The normative aspect of AI Alignment seems to be under researched.</p><p><strong>Dylan:</strong> Can you say a little more about what you mean by that?</p><p><strong>Lucas:</strong> What I meant was the normative deliberative process, the difficulty in coming to normative conclusions and what the appropriate epistemic and deliberative process is for arriving at normative solutions and how narrow AI systems can take us to a beautiful world where advanced AI systems actually lead us to post human ethics.</p><p>If we ever want to get to a place where general systems take us to post human ethics, why not start today with figuring out how narrow systems can work to amplify human moral decision making and deliberative processes.</p><p><strong>Dylan: </strong>I think the hard part there is, I don’t exactly know what it means to amplify those processes. My perspective is that we as a species do not yet have a good understanding of what those deliberative processes actually represent and what formed the result actually does.</p><p><strong>Lucas: </strong>It’s just like giving more information, providing tons of data, analyzing the data, potentially pointing out biases. The part where they’re literally amplifying cognitive implicit or explicit decision making process is more complicated and will require more advancement and cognition and deliberation and stuff. But yeah, I still think there are more mundane ways in which it can make us better moral reasoners and decision makers.</p><p>If I could give you like 10,000 more bits of information every day about moral decisions that you make, you would probably just be a better moral agent.</p><p><strong>Dylan:</strong> Yes, one way to try to think about that is maybe things like VR approaches to increasing empathy. I think that that has a lot of power to make us better.</p><p><strong>Lucas: </strong>Max always says that there’s a race between wisdom and the power of our technology and it seems like people really aren’t taking seriously ways in which we can amplify wisdom because wisdom is generally taken to be part of the humanities and like the soft sciences. Maybe we should be taking more seriously ways in which narrow current day AI systems can be used to amplify the progress at which the human species makes wisdom. Because otherwise we’re just gonna like continue how we always continue and the wisdom is going to go really slow and then we’re going to probably learn from a bunch of mistakes.</p><p>And it’s just not going to be as good until we’ll develop a rigorous science of making moral progress or like using technology to amplify the progress of wisdom and moral progress.</p><p><strong>Dylan: </strong>So in principle, what you’re saying, I don’t really disagree with it, but I also don’t know how that would change what I’m working on either. In the sense that I’m not sure what it would mean. I do not know how I would do research on amplifying wisdom. I just don’t really know what that means. And that’s not to say it’s an impossible problem, we talked earlier about how I don’t know what partial value alignment means, that something that you and I can talk about it and we can intuitively I think align on a concept, but it’s not a concept I knew how to translate into actionable concrete research problems right now.</p><p>In the same way, the idea of amplifying wisdom and making people more wise is something that I think intuitively I understand what you mean, but when I try to think about how an AI system would make someone wiser, that feels difficult.</p><p><strong>Lucas: </strong>It can seem difficult, but I feel like it would, obviously this is like an open research question, but if you were able to identify a bunch of variables that are most important for moral decision making and then if you could use AI systems to sort of gather aggregate and compile in certain ways and analyze moral information in this way, again, it just seems more tractable than people seem to be letting on.</p><p><strong>Dylan: </strong>Yeah, although I wonder now is that different from value alignment does, we’re thinking about it, right? Concrete research thing I spend a while thinking about is, how do you identify the features that a person considers to be valuable? Say, we don’t know the relative tradeoffs between them.</p><p>One way you might try to solve value alignment is have a process that identifies the features that might matter in the world and then have a second process that identifies the appropriate tradeoffs between those features, and maybe something about diminishing returns or something like that. And that to me sounds like I just placed values with wisdom and I’ve got sort of what you’re thinking about. I think both of those terms are similarly diffuse. I wonder if what we’re talking about is semantics, and if it’s not, I’d like to know what the difference is.</p><p><strong>Lucas:</strong> I guess, the more mundane definition of wisdom, at least in the way that Max Tegmark would use it would be like the ways in which we use our technology. I might have specific preferences, but just because I have specific preferences that I may or may not be aligning an AI system to does not necessarily mean that that total process, this like CIRL process is actually an expression of wisdom.</p><p><strong>Dylan: </strong>Okay, can you provide a positive description of what a process would look like? Or like basically what I’m saying is I can hear the point of I have preferences and I aligned my system to it and that’s not necessarily a wise system and …</p><p><strong>Lucas: </strong>Yeah, like I build a fire because I want to be hot, but then the fire catches my village on fire and no longer is … That’s still might be value alignment.</p><p><strong>Dylan: </strong>But isn’t [crosstalk 00:48:39] some values that you didn’t take into account when you were deciding to build the fire.</p><p><strong>Lucas: </strong>Yeah, that’s right. So I don’t know. I’d probably have to think about this more because I guess this is something that I just sort of throwing out right now as a reaction to what we’ve been talking about. So I don’t have a very good theory of it.</p><p><strong>Dylan: </strong>And I don’t wanna say that you need to know the right answers to these things to not have that be a useful direction to push people.</p><p><strong>Lucas: </strong>We don’t want to use different concepts to just reframe the same problem and just make a conceptual mess.</p><p><strong>Dylan: </strong>That’s what I’m a little bit concerned about and that’s the thing I’m concerned about broadly. We’ve got a lot of issues that we’re thinking about in dealing with that we’re not really sure what they are.</p><p>For me, I think one of the really helpful things has been to frame the issue that I’m thinking about as if a person has a behavior that they want to implement into the world and that’s a complex behavior that they don’t know how to identify immediately. How do you actually go about building systems that allow you to implement that behavior effectively, evaluate that the behavior is actually been correctly implemented.</p><p><strong>Lucas: </strong>Avoiding side effects, avoiding …</p><p><strong>Dylan:</strong> Like all of these kinds of things that we sort of concerned about in AI Safety, in my mind fall a bit more into place when we frame the problem as I have a desired behavior that I want to exist, a response function, a policy function that I want to implement into the world. What are the technological systems I can use to implement that in a computer or a robot or what have you.</p><p><strong>Lucas: </strong>Okay. Well, do you have anything else you’d like to wrap up on?</p><p><strong>Dylan:</strong> No, I just, I want to say thanks for asking hard questions and making me feel uncomfortable because I think it’s important to do a lot of that as a scientist and in particular I think as people working on AI, we should be spending a bit more time being uncomfortable and talking about these things, because it does impact what we end up doing and it does I think impact the trajectories that we put the technology on.</p><p><strong>Lucas: </strong>Wonderful. So if people want to read about cooperative inverse reinforcement learning, where can we find the paper or other work that you have on that? What do you think are the best resources? What are just general things you’d like to point people towards in order to follow you or keep up to date with AI Alignment?</p><p><strong>Dylan: </strong>I tweet occasionally about AI Alignment and a bit of AI ethics questions, the Hadfield-Menell, my first initial, last name. And if you’re interested in getting a technical introduction to value alignment, I would say take a look at the 2016 paper on cooperative IRL. If you’d like a more general introduction, there’s a blog post from summer 2017 on the bear blog.</p><p><strong>Lucas:</strong> All right, thanks so much Dylan, and maybe we’ll be sitting in a similar room again in two years for Beneficial Artificial Super Intelligence 2021.</p><p><strong>Dylan:</strong> I look forward to it. Thanks a bunch.</p><p><strong>Lucas:</strong> Thanks. See you, Dylan. If you enjoyed this podcast, please subscribe, give it a like, or share it on your preferred social media platform. We’ll be back again soon with another episode in the AI Alignment series.</p><p><em>[end of recorded material]</em></p><p><em>Originally published at </em><a href="https://futureoflife.org/2019/01/17/cooperative-inverse-reinforcement-learning-with-dylan-hadfield-menell/"><em>futureoflife.org</em></a><em> on January 17, 2019.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bbbf831740c0" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Podcast: Existential Hope in 2019 and Beyond]]></title>
            <link>https://medium.com/@FLIxrisk/podcast-existential-hope-in-2019-and-beyond-aa732d59bcaf?source=rss-e33e2d2a809c------2</link>
            <guid isPermaLink="false">https://medium.com/p/aa732d59bcaf</guid>
            <category><![CDATA[cosmos]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[existential-risk]]></category>
            <category><![CDATA[hope]]></category>
            <dc:creator><![CDATA[Future of Life]]></dc:creator>
            <pubDate>Fri, 21 Dec 2018 22:09:40 GMT</pubDate>
            <atom:updated>2019-01-18T16:45:33.138Z</atom:updated>
            <content:encoded><![CDATA[<iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fw.soundcloud.com%2Fplayer%2F%3Furl%3Dhttps%253A%252F%252Fapi.soundcloud.com%252Ftracks%252F548436468%26show_artwork%3Dtrue&amp;url=https%3A%2F%2Fsoundcloud.com%2Ffutureoflife%2Fexistential-hope-in-2019-and-beyond&amp;image=http%3A%2F%2Fi1.sndcdn.com%2Fartworks-000462174552-biqdbo-t500x500.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=soundcloud" width="800" height="166" frameborder="0" scrolling="no"><a href="https://medium.com/media/13904e50ec4225c40b1a00484ad141af/href">https://medium.com/media/13904e50ec4225c40b1a00484ad141af/href</a></iframe><p>Humanity is at a turning point. For the first time in history, we have the technology to completely obliterate ourselves. But we’ve also created boundless possibilities for all life that could enable just about any brilliant future we can imagine. Humanity could erase itself with a nuclear war or a poorly designed AI, or we could colonize space and expand life throughout the universe: As a species, our future has never been more open-ended.</p><p>The potential for disaster is often more visible than the potential for triumph, so as we prepare for 2019, we want to talk about existential hope, and why we should actually be more excited than ever about the future. In this podcast, Ariel talks to six experts–Anthony Aguirre, Max Tegmark, Gaia Dempsey, Allison Duettmann, Josh Clark, and Anders Sandberg–about their views on the present, the future, and the path between them.</p><p><a href="http://www.anthony-aguirre.com/">Anthony</a> and <a href="https://space.mit.edu/home/tegmark/">Max</a> are both physics professors and cofounders of FLI. <a href="https://futureofstorytelling.org/speaker/gaia-dempsey">Gaia</a> is a tech enthusiast and entrepreneur, and with her newest venture, <a href="http://7thfuture.com/">7th Future</a>, she’s focusing on bringing people and organizations together to imagine and figure out how to build a better future. <a href="https://foresight.org/our-team/allison-duettmann/">Allison</a> is a researcher and program coordinator at the <a href="https://foresight.org/">Foresight Institute</a> and creator of the website <a href="https://www.existentialhope.com/">existentialhope.com</a>. <a href="https://www.stuffyoushouldknow.com/about-josh.htm">Josh</a> is cohost on the <a href="https://www.stuffyoushouldknow.com/podcasts">Stuff You Should Know Podcast</a>, and he recently released a 10-part series on existential risks called <a href="https://www.theendwithjosh.com/">The End of the World with Josh Clark</a>. <a href="https://www.fhi.ox.ac.uk/team/anders-sandberg/">Anders</a> is a senior researcher at the <a href="https://www.fhi.ox.ac.uk/">Future of Humanity Institute</a> with a background in computational neuroscience, and for the past 20 years, he’s studied the ethics of human enhancement, existential risks, emerging technology, and life in the far future.</p><p>We hope you’ll come away feeling inspired and motivated–not just to prevent catastrophe, but to facilitate greatness.</p><p>Topics discussed in this episode include:</p><ul><li>How technology aids us in realizing personal and societal goals.</li><li>FLI’s successes in 2018 and our goals for 2019.</li><li>Worldbuilding and how to conceptualize the future.</li><li>The possibility of other life in the universe and its implications for the future of humanity.</li><li>How we can improve as a species and strategies for doing so.</li><li>The importance of a shared positive vision for the future, what that vision might look like, and how a shared vision can still represent a wide enough set of values and goals to cover the billions of people alive today and in the future.</li><li>Existential hope and what it looks like now and far into the future.</li></ul><p>You can listen to the podcast above, or read the full transcript below.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*B8qUtYVS34PKVp5t" /></figure><p><strong>Ariel:</strong> Hi everyone. Welcome back to the FLI podcast. I’m your host, Ariel Conn, and I am truly excited to bring you today’s show. This month, we’re departing from our standard two-guest interview format because we wanted to tackle a big and fantastic topic for the end of the year that would require insight from a few extra people. It may seem as if we at FLI spend a lot of our time worrying about existential risks, but it’s helpful to remember that we don’t do this because we think the world will end tragically: We address issues relating to existential risks because we’re so confident that if we can overcome these threats, we can achieve a future greater than any of us can imagine.</p><p>And so, as we end 2018 and look toward 2019, we want to focus on a message of hope, a message of <em>existential</em> hope.</p><p>I’m delighted to present Anthony Aguirre, Max Tegmark, Gaia Dempsey, Allison Duettmann, Josh Clark and Anders Sandberg, all of whom were kind enough to come on the show and talk about why they’re so hopeful for the future and just how amazing that future could be.</p><p>Anthony and Max are both physics professors and cofounders of FLI. Gaia is a tech enthusiast and entrepreneur, and with her newest venture, 7th Future, she’s focusing on bringing people and organizations together to imagine and figure out how to build a better future. Allison is a researcher and program coordinator at the Foresight Institute and she created the website existentialhope.com. Josh is cohost on the Stuff You Should Know Podcast, and he recently released a 10-part series on existential risks called The End of the World with Josh Clark. Anders is a senior researcher at the Future of Humanity Institute with a background in computational neuroscience, and for the past 20 years, he’s studied the ethics of human enhancement, existential risks, emerging technology, and life in the far future.</p><p>Over the course of a few days, I interviewed all six of our guests, and I have to say, it had an incredibly powerful and positive impact on my psyche. We’ve merged these interviews together for you here, and I hope you’ll all also walk away feeling a bit more hope for humanity’s collective future, whatever that might be.</p><p>But before we go too far into the future, let’s start with Anthony and Max, who can talk a bit about where we are today.</p><p><strong>Anthony:</strong> I’m Anthony Aguirre, I’m one of the founders of the Future of Life Institute. And in my day job, I’m a Physicist at the University of California at Santa Cruz.</p><p><strong>Max:</strong> I am Max Tegmark, a professor doing physics and AI research here at MIT, and also the president of the Future of Life Institute.</p><p><strong>Ariel:</strong> All right. Thank you so much for joining us today. I’m going to start with sort of a big question. That is, do you think we can use technology to solve today’s problems?</p><p><strong>Anthony:</strong> I think we can use technology to solve any problem in the sense that I think technology is an extension of our capability: it’s something that we develop in order to accomplish our goals and to bring our will into fruition. So, sort of by definition, when we have goals that we want to do — problems that we want to solve — technology should in principle be part of the solution.</p><p><strong>Max:</strong> Take, for example, poverty. It’s not like we don’t have the technology right now to eliminate poverty. But we’re steering the technology in such a way that there are people who starve to death, and even in America there are a lot of children who just don’t get enough to eat, through no fault of their own.</p><p><strong>Anthony:</strong> So I’m broadly optimistic that, as it has over and over again, technology will let us do things that we want to do better than we were previously able to do them. Now, that being said, there are things that are more amenable to better technology, and things that are less amenable. And there are technologies that tend to, rather than functioning as kind of an extension of our will, will take on a bit of a life of their own. If you think about technologies like medicine, or good farming techniques, those tend to be sort of overall beneficial and really are kind of accomplishing purposes that we set. You know, we want to be more healthy, we want to be better fed, we build the technology and it happens. On the other hand, there are obviously technologies that are just as useful or even more useful for negative purposes — socially negative or things that most people agree are negative things: landmines, for example, as opposed to vaccines. These technologies come into being because somebody is trying to accomplish their purpose — defending their country against an invading force, say — but once that technology exists, it’s kind of something that is easily used for ill purposes.</p><p><strong>Max:</strong> Technology simply empowers us to do good things or bad things. Technology isn’t evil, but it’s also not good. It’s morally neutral. Right? You can use fire to warm up your home in the winter or to burn down your neighbor’s house. We have to figure out how to steer it and where we want to go with it. I feel that there’s been so much focus on just making our tech powerful right now — because that makes money, and it’s cool — that we’ve neglected the steering and the destination quite a bit. And in fact, I see the core goal of the Future of Life Institute: Help bring back focus on the steering of our technology and the destination.</p><p><strong>Anthony:</strong> There are also technologies that are really tricky in that they give us what we think we want, but then we sort of regret having later, like addictive drugs, or gambling, or cheap sugary foods, or-</p><p><strong>Ariel:</strong> Social media.</p><p><strong>Anthony:</strong> … certain online platforms that will go unnamed. We feel like this is what we want to do at the time; We choose to do it. We choose to eat the huge sugary thing, or to spend some time surfing the web. But later, with a different perspective maybe, we look back and say, “Boy, I could’ve used those calories, or minutes, or whatever, better.” So who’s right? Is it the person at the time who’s choosing to eat or play or whatever? Or is it the person later who’s deciding, “Yeah, that wasn’t a good use of my time or not.” Those technologies I think are very tricky, because in some sense they’re giving us what we want. So we reward them, we buy them, we spend money, the industries develop, the technologies have money behind them. At the same time, it’s not clear that they make us happier.</p><p>So I think there are certain social problems, and problems in general, that technology will be tremendously helpful in improving as long as we can act to sort of wisely try to balance the effects of technology that have dual use toward the positive, and as long as we can somehow get some perspective on what to do about these technologies that take on a life of their own, and tend to make us less happy, even though we dump lots of time and money into them.</p><p><strong>Ariel:</strong> This sort of idea of technologies — that we’re using them and as we use them we think they make us happy and then in the long run we sort of question that — is this a relatively modern problem, or are there examples of anything that goes further back that we can learn from from history?</p><p><strong>Anthony:</strong> I think it goes fairly far back. Certainly drug use goes a fair ways back. I think there have been periods where drugs were used as part of religious or social ceremonies and in other kind of more socially constructive ways. But then, it’s been a fair amount of time where opiates and very addictive things have existed also. Those have certainly caused social problems back at least a few centuries.</p><p>I think a lot of these examples of technologies that give us what we seem to want but not really what we want are ones in which we’re applying the technology to a species — us — that developed in a very different set of circumstances, and that contrast between what’s available and what we evolutionarily wanted is causing a lot of problems. The sugary foods are an obvious example where we can now just supply huge plenitudes of something that was very rare and precious back in more evolutionary times — you know, sweet calories.</p><p>Drugs are something similar. We have a set of chemistry that helps us out in various situations, and then we’re just feeding those same chemical pathways to make ourselves feel good in a way that is destructive. And violence might be something similar. Violent technologies go way, way back. Those are another one that are clearly things that we want to invent to further our will and accomplish our goals. They’re also things that may at some level be addictive to humans. I think it’s not entirely clear exactly how — there’s a strange mix there, but I think there’s certainly something compelling and built into at least many humans’ DNA that promotes fighting and hunting and all kinds of things that were evolutionarily useful way back when and perhaps less useful now. It had a clear evolutionary purpose with tribes that had to defend themselves, with animals that needed to be killed for food. But feeding that desire to run around and hunt and shoot people, which most people aren’t doing in real life, but tons of people are doing in video games. So there’s clearly some built in mechanism that’s rewarding that behavior as being fun to do and compelling. Video games are obviously a better way to express that than running around and doing it in real life, but it tells you something about some circuitry that is still there and is left over from early times. So I think there are a number of examples like that — this connection between our biological evolutionary history and what technology makes available in large quantities — where we really have to think carefully about how we want to play that.</p><p><strong>Ariel:</strong> So, as you look forward to the future, and sort of considering some of these issues that you’ve brought up, how do you envision us being able to use technology for good and maybe try to overcome some of these issues? I mean, maybe it is good if we’ve got people playing video games instead of going around shooting people in real life.</p><p><strong>Anthony:</strong> Yeah. So there may be examples where some of that technology can fulfill a need in a less destructive way than it might otherwise be. I think there are also plenty of examples where a technology can root out or sort of change the nature of a problem that would be enormously difficult to do something about without a technology. So for example, I think eating meat, when you analyze it from almost any perspective, is a pretty destructive thing for humanity to be doing. Ecologically, ethically in terms of the happiness of the animals, health-wise: so many things are destructive about it. And yet, you really have the sense that it’s going to be enormously difficult — it would be very unlikely for that to change wholesale on a relatively short period of time.</p><p>However, there are technologies — clean meat, cultured meat, really good tasting vegetarian meat substitutes — that are rapidly coming to market. And you could imagine if those things were to get cheap and widely available and perhaps a little bit healthier, that could dramatically change that situation relatively quickly. I think if a non-ecologically destructive, non-suffering inducing, just as tasty and even healthier product were cheaper, I don’t think people would be eating meat. Very few people actually like, I think, intrinsically the idea of having an animal suffer in order for them to eat. So I think that’s an example of something that would be really, really hard to change through just social actions. Could be jump started quite a lot by technology — that’s one of the ones I’m actually quite hopeful about.</p><p>Global warming I think is a similar one — it’s on some level a social and economic problem. It’s a long-term planning problem, which we’re very bad at. It’s pretty clear how to solve the global warming issue if we really could think on the right time scales and weigh the economic costs and benefits over decades — it’d be quite clear that mitigating global warming now and doing things about it now might take some overall investment that would clearly pay itself off. But we seem unable to accomplish that.</p><p>On the other hand, you could easily imagine a really cheap, really power-dense, quickly rechargeable battery being invented and just utterly transforming that problem into a much, much more tractable one. Or feasible, small-scale nuclear fusion power generation that was cheap. You can imagine technologies that would just make that problem so much easier, even though it is ultimately kind of a social or political problem that could be solved. The technology would just make it dramatically easier to do that.</p><p><strong>Ariel:</strong> Excellent. And so thinking more hopefully — even when we’re looking at what’s happening in the world today, news is usually focusing on all the bad things that have gone wrong — when you look around the world today, what do you think, “Wow, technology has really helped us achieve this, and this is super exciting?”</p><p><strong>Max:</strong> Almost everything I love about today is the result of technology. It’s because of technology that we’ve more than doubled the lifespan that we humans used to have, most of human history. More broadly, I feel that the technology is empowering us. Ten thousand years ago, we felt really, really powerless; We were these beings, you know, looking at this great world out there and having very little clue about how it worked — it was largely mysterious to us — and even less ability to actually influence the world in a major way. Then technology enabled science, and vice versa. So the sciences let us understand more and more how the world works, and let us build this technology which lets us shape the world to better suit us. Helping produce much better, much more food, helping keep us warm in the winter, helping make hospitals that can take care of us, and schools that can educate us, and so on.</p><p><strong>Ariel:</strong> Let’s bring on some of our other guests now. We’ll turn first to Gaia Dempsey. How do you envision technology being used for good?</p><p><strong>Gaia:</strong> That’s a huge question.</p><p><strong>Ariel:</strong> It is. Yes.</p><p><strong>Gaia:</strong> I mean, at its essence I think technology really just means a tool. It means a new way of doing something. Tools can be used to do a lot of good — making our lives easier, saving us time, helping us become more of who we want to be. And I think technology is best used when it supports our individual development in the direction that we actually want to go — when it supports our deeper interests and not just the, say, commercial interests of the company that made it. And I think in order for that to happen, we need for our society to be more literate in technology. And to me that’s not just about understanding how computing platforms work, but also understanding the impact that tools have on us as human beings. Because they don’t just shape our behavior, they actually shape our minds and how we think.</p><p>So I think we need to be very intentional about the tools that we choose to use in our own lives, and also the tools that we build as technologists. I’ve always been very inspired by Douglas Engelbart’s work, and I think that — I was revisiting his original conceptual framework on augmenting human intelligence, which he wrote and published in 1962 — and I really think he had the right idea, which is that tools used by human beings don’t exist in a vacuum. They exist in a coherent system and that system involves language: the language that we use to describe the tools and understand how we’re using them; the methodology; and of course the training and education around how we learn to use those tools. And I think that as a tool maker it’s really important to think about each of those pieces of an overarching coherent system, and imagine how they’re all going to work together and fit into an individual’s life and beyond: you know, the level of a community and a society.</p><p><strong>Ariel:</strong> I want to expand on some of this just a little bit. You mentioned this idea of making sure that the tool, the technology tool, is being used for people and not just for the benefit, the profit, of the company. And that that’s closely connected to making sure that people are literate about the technology. One, just to confirm that that is actually what you were saying. And, two, I mean one of the reasons I want to confirm this is because that is my own concern — that it’s being too focused for making profit and not enough people really understand what’s happening. My question to you is, then, how do we educate people? How do we get them more involved?</p><p><strong>Gaia:</strong> I think for me, my favorite types of tools are the kinds of tools that support us in developing our thinking and that help us accelerate our ability to learn. But I think that some of how we do this in our society is not just about creating new tools or getting trained on new tools, but really doesn’t have very much to do with technology at all. And that’s in our education system, teaching critical thinking. And teaching, starting at a young age, to not just accept information that is given to you wholesale, but really to examine the motivations and intentions and interests of the creator of that information, and the distributor of that information. And I think these are really just basic tools that we need as citizens in a technological society and in a democracy.</p><p><strong>Ariel:</strong> That actually moves nicely to another question that I have. Well, I actually think the sentiment might be not quite as strong as it once was, but I do still hear a lot of people who sort of approach technology as the solution to any of today’s problems. And I’m personally a little bit skeptical that we can only use technology. I think, again, it comes back to what you were talking about with it’s a tool so we can use it, but I think it just seems like there’s more that needs to be involved. I guess, how do you envision using technology as a tool, and still incorporating some of these other aspects like teaching critical thinking?</p><p><strong>Gaia:</strong> You’re really hitting on sort of the core questions that are fundamental to creating the kind of society that we want to live in. And I think that we would do well to spend more time thinking deeply about these questions. I think technology can do really incredible, tremendous things in helping us solve problems and create new capabilities. But it also creates a new set of problems for us to engage with.</p><p>We’ve sort of coevolved with our technology. So it’s easy to point to things in the culture and say, “Well, this never would have happened without technology X.” And I think that’s true for things that are both good and bad. I think, again, it’s about taking a step back and taking a broader view, and really not just teaching critical thinking and critical analysis, but also systems level thinking. And understanding that we ourselves are complex systems, and we’re not perfect in the way that we perceive reality — we have cognitive biases, we cannot necessarily always trust our own perceptions. And I think that’s a lifelong piece of work that everyone can engage with, which is really about understanding yourself first. This is something that Yuval Noah Harari talked about in a couple of his recent books and articles that he’s been writing, which is: if we don’t do the work to really understand ourselves first and our own motivations and interests, and sort of where we want to go in the world, we’re much more easily co-opted and hackable by systems that are external to us.</p><p>There are many examples of recommendation algorithms and sentiment analysis — audience segmentation tools that companies are using to be able to predict what we want and present that information to us before we’ve had a chance to imagine that that is something we could want. And while that’s potentially useful and lucrative for marketers, the question is what happens when those tools are then utilized not just to sell us a better toothbrush on Amazon, but when it’s actually used in a political context. And so with the advent of these vast machine learning, reinforcement learning systems that can look at data and look at our behavior patterns and understand trends in our behavior and our interests, that presents a really huge issue if we are not ourselves able to pause and create a gap, and create a space between the information that’s being presented to us within the systems that we’re utilizing and really our own internal compass.</p><p><strong>Ariel:</strong> You’ve said two things that I think are sort of interesting, especially when they’re brought together. And the first is this idea that we’ve coevolved with technology — which, I actually hadn’t thought of it in that phrase before, and I think it’s a really, really good description. But then when we consider that we’ve coevolved with technology, what does that mean in terms of knowing ourselves? And especially knowing ourselves as our biological bodies, and our limiting cognitive biases? I don’t know if that’s something that you’ve thought about much, but I think that combination of ideas is an interesting one.</p><p><strong>Gaia:</strong> I mean, I know that I certainly already feel like I’m a cyborg. Part of knowing myself is — it does involve understanding the tools that I use, that feel that they are extensions of myself. That kind of comes back to the idea of technology literacy, and systems literacy, and being intentional about the kinds of tools that I want to use. For me, my favorite types of tools are the kind that I think are very rare: the kind that support us developing the capacity for long-term thinking, and for being true to the long-term intentions and goals that I set for myself.</p><p><strong>Ariel:</strong> Can you give some examples of those?</p><p><strong>Gaia:</strong> Yeah, I’ll give a couple examples. One example that’s sort of probably familiar to a lot of people listening to this comes from the book Ready Player One. And in this book the main character is interacting with his VR system that he sort of lives and breathes in every single day. And at a certain point the system asks him: do you want to activate your health module? I forgot exactly what it was called. And without giving it too much thought, he kind of goes, “Sure. Yeah, I’d like to be healthier.” And it instantiates a process whereby he’s not allowed to log into the OASIS without going through his exercise routine every morning. To me, what’s happening there is: there is a choice.</p><p>And it’s an interesting system design because he didn’t actually do that much deep thinking about, “Oh yeah, this is a choice I really want to commit to.” But the system is sort of saying, “We’re thinking through the way that your decision making process works, and we think that this is something you really do want to consider. And we think that you’re going to need about three months before you make a final decision as to whether this is something you want to continue with.”</p><p>So that three month period or whatever, and I believe it was three months in the book, is what’s known as an akrasia horizon. Which is a term that I learned through a different tool that is sort of a real life version of that, which is called Beeminder. And the akrasia horizon is, really, it’s a time period that’s long enough that it will sort of circumvent a cognitive bias that we have to really prioritize the near term at the expense of the future. And in the case of the Ready Player One example, the near term desire that he would have that would circumvent the future — his long-term health — is, “I don’t feel like working out today. I just want to get into my email or I just want to play a video game right now.” And a very similar sort of setup is created in this tool Beeminder, which I love to use to support some goals that I want to make sure I’m really very motivated to meet.</p><p>So it’s a tool where you can put in your goals and you can track them either yourself by entering the data manually, or you can connect to a number of different tracking capabilities like RescueTime and others. And if you don’t stay on track with your goals, they charge your credit card. It’s a very effective sort of motivating force. And so I sort of have a nickname: I call these systems time bridges. Which are really choices made by your long-term thinking self, that in some way supersedes the gravitational pull toward mediocrity inherent in your short-term impulses.</p><p>It’s about experimenting too. And this is one particular system that creates consequences and accountability. And I love systems. For me if I don’t have systems in my life that help me organize the work that I want to do, I’m hopeless. That’s why I like to collect and I’m sort of an avid taster of different systems, and I’ll try anything, and really collect and see what works. And I think that’s important. It’s a process of experimentation to see what works for you.</p><p><strong>Ariel:</strong> Let’s turn to Allison Duettmann now, for her take on how we can use technology to help us become better versions of ourselves and to improve our societal interactions.</p><p><strong>Allison:</strong> I think there are a lot of technological tools that we can use to aid our reasoning and sense making and coordination. So I think that technologies can be used to help with reasoning, for example, by mitigating trauma, or bias, or by augmenting our intelligence. That’s the whole point of creating AI in the first place. Technologies can also be used to help with collective sense-making, for example with truth-finding and knowledge management, and I think your hypertexts and prediction markets — something that Anthony’s working on — are really worthy examples here. I also think technologies can be used to help with coordination. Mark Miller, who I’m currently writing a book with, likes to say that if you lower the risks of cooperation, you’ll get a more cooperative world. I think that most cooperative interactions may soon be digital.</p><p><strong>Ariel:</strong> That’s sort of an interesting idea, that there’s risks to cooperation. Can you maybe expand on that a little bit more?</p><p><strong>Allison:</strong> Yeah, sure. I think that most of our interactions are already digital ones, for some of us at least, and they will be more and more so in the future. So I think that one step to lowering the risk of cooperation is establishing cybersecurity as a first step, because this would decrease the risk of digital coercion. But I do think that’s only part of it, because rather than just freeing us from the restraints that keep us from cooperating, we also need to equip us with the tools to cooperate, right?</p><p><strong>Ariel:</strong> Yes.</p><p><strong>Allison:</strong> I think some of those may be smart contracts to allow individuals to credibly commit, but there may be others too. I just think that we have to realize that the same technologies that we’re worried about in terms of risks are also the ones that may augment our abilities to decrease those risks.</p><p><strong>Ariel:</strong> One of the things that came to mind as you were talking about this, using technology to improve cooperation — when we look at the world today, technology isn’t spread across the globe evenly. People don’t have equal access to these tools that could help. Do you have ideas for how we address various inequality issues, I guess?</p><p><strong>Allison:</strong> I think inequality is a hot topic to address. I’m currently writing a book with Mark Miller and Christine Peterson on a few strategies to strengthen civilization. In this book we outline a few paths to do so, but also potential positive outcomes. One of the outcomes that we’re outlining is a voluntary world in which all entities can cooperate freely with each other to realize their interests. It’s kind of based on the premise that finding one utopia that works for everyone is hard, and is perhaps impossible, but that in the absence of knowing what’s in everyone’s interest, we shouldn’t try to impose any interests by one entity — whether that’s an AI or an organization or a state — but we should try to create a framework in which different entities, with different interests, whether they’re human or artificial, can pursue their interests freely by cooperating. And I think If you look at the strategy, it has worked pretty well so far. If you look at society right now it’s really not perfect, but by allowing humans to cooperate freely and engage in some mutually beneficial relationships, civilization already serves our interests quite well. And it’s really not perfect by far, I’m not saying this, but I think as a whole, our civilization at least tends imperfectly to plan for pareto-preferred paths. We have survived so far, and in better and better ways.</p><p>So a few ways that we propose to strengthen this highly involved process is by proposing kind of general recommendations for solving coordination problems, and then a few more specific ideas on reframing a few risks. But I do think that enabling a voluntary world in which different entities can cooperate freely with each other is the best we can do, given our limited knowledge of what is in everyone’s interests.</p><p><strong>Ariel:</strong> I find that interesting, because I hear lots of people focus on how great intelligence is, and intelligence is great, but it does often seem — and I hear other people say this — that cooperation is also one of the things that our species has gotten right. We fail at it sometimes, but it’s been one of the things, I think, that’s helped.</p><p><strong>Allison:</strong> Yeah, I agree. I hosted an event last year at the Internet Archive on different definitions of intelligence. Because in the paper that we wrote last year, we have this very grand, or broad conception of intelligence, which includes civilization as an intelligence. So I think you may be asking yourself the question of, what does it mean to be intelligent, and if what we care about is problem-solving ability then I think that civilization certainly classifies as a system that can solve more problems than any individual that is within it alone. So I do think this is part of the cooperative nature of the individual parts within civilization, and so I don’t think that cooperation and intelligence are mutually exclusive at all. Marvin Minsky wrote this amazing book, Society of Mind, and in much of this, has similar ideas.</p><p><strong>Ariel:</strong> I’d like to take this idea and turn it around, and this is a question specifically for Max and Anthony: looking back at this past year, how has FLI helped foster cooperation and public engagement surrounding the issues we’re concerned about? What would you say were FLI’s greatest successes in 2018?</p><p><strong>Anthony:</strong> Let’s see, 2018. What I’ve personally enjoyed the most, I would say, is starting the engagement between the technical researchers and the nonprofit community really starting to get more engaged with state and federal governments. So for example the Asilomar principles — which were generated at this nexus of business and nonprofit and academic thinkers about AI and related things — I think were great. But that conversation didn’t really include much from people in policy, and governance, and governments, and so on. So, starting to see that thinking, and those recommendations, and those aspirations of the community of people who know about AI and are thinking hard about it and what it should do and what it shouldn’t do — seeing that start to come into the political sphere, and the government sphere, and the policy sphere I think is really encouraging.</p><p>That seems to be happening in many places at some level. I think the local one that I’m excited about is the passage of the California legislature of a resolution endorsing the Asilomar principles. That felt really good to see that happen and really encouraging that there were people in the legislature that — we didn’t go and lobby them to do that, they came to us and said, “This is really important. We want to do something.” And we worked with them to do that. That was super encouraging, because it really made it feel like there is a really open door, and there’s a desire in the policy world to do something. This thing is getting on people’s radar, that there’s a huge transformation coming from AI.</p><p>They see that their responsibility is to do something about that. They don’t intrinsically know what they should be doing, they’re not experts in AI, they haven’t been following the field. So there needs to be that connection and it’s really encouraging to see how open they are and how much can be produced with honestly not a huge level of effort; Just communication and talking through things I think made a significant impact. I was also happy to see how much support there continues to be for controlling the possibility of lethal autonomous weapons.</p><p>The thing we’ve done this year, the lethal autonomous weapons pledge, I felt really good about the success of. So this was an idea that anybody who’s interested, but especially companies who are engaged in developing related technologies, drones, or facial recognition, or robotics, or AI in general — to get them to take that step themselves of saying, “No, we want to develop these technologies for good, and we have no interest in developing things that are going to be weaponized and used in lethal autonomous weapons.”</p><p>I think having a large number of people and corporations sign on to a pledge like that is useful not so much because they were planning to do all those things and now they signed a pledge, so they’re not going to do it anymore. I think that’s not really the model so much as it’s creating a social and cultural norm that these are things that people just don’t want to have anything to do with, just like biotech companies don’t really want to be developing biological weapons, they want to be seen as forces for good that are building medicines and therapies and treatments and things. Everybody is happy for biotech companies to be doing those things.</p><p>If biotech companies were building biological weapons also, you really start to wonder, “Okay, wait a minute, why are we supporting this? What are they doing with my information? What are they doing with all this genetics that they’re getting? What are they doing with the research that’s funded by the government? Do we really want to be supporting this?” So keeping that distinction in the industry between all the things that we all support — better technologies for helping people — versus the military applications, particularly in this rather destabilizing and destructive way: I think that is more the purpose — to really make clear that there are companies that are going to develop weapons for the military, and that’s part of the reality of the world.</p><p>We have militaries; We need, at the moment, militaries. I think I certainly would not advocate that the US should stop defending itself, or shouldn’t develop weapons, and I think it’s good that there are companies that are building those things. But there are very tricky issues when the companies building military weapons are the same companies that are handling all of the data of all of the people in the world or in the country. I think that really requires a lot of thought, how we’re going to handle it. And seeing companies engage with those questions and thinking about how are the technologies that we’re developing, how are they going to be used and for what purposes, and what purposes do we not want them to be used for is really, really heartening. It’s been very positive I think to see at least in certain companies those sort of conversations go on with our pledge or just in other ways.</p><p>You know, seeing companies come out with, “This is something that we’re really worried about. We’re developing these technologies, but we see that there could be major problems with them.” That’s very encouraging. I don’t think it’s necessarily a substitute for something happening at the regulatory or policy level, I think that’s probably necessary too, but it’s hugely encouraging to see companies being proactive about thinking about the societal and ethical implications of the technologies they’re developing.</p><p><strong>Max:</strong> There are four things I’m quite excited about. One of them is that we managed to get so many leading companies and AI researchers and universities to pledge to not build lethal autonomous weapons, also known as killer robots. Second is that we were able to channel two million dollars, thanks to Elon Musk, to 10 research groups around the world to help figure out how to make artificial general intelligence safe and beneficial. Third is that the state of California decided to officially endorse the 23 Asilomar Principles. It’s really cool that these are getting more taken seriously now, even by policy makers. And the fourth is that we were able to track down the children of Stanislav Petrov in Russia, thanks to whom this year is not the 35th anniversary year of World War III, and actually give them the appreciation we feel that they deserve.</p><p>I’ll tell you a little more about this one because it’s something I think a lot of people still aren’t that aware of. But September 26th, 35 years ago, Stanislav Petrov was on shift and in charge of his Soviet early warning station, which showed five US nuclear missiles incoming, one after the other. Obviously, not what he was hoping that would happen at work that day and a really horribly scary situation where the natural response is to do what that system was built for: namely, warning the Soviet Union so that they would immediately strike back. And if that had happened, then thousands of mushroom clouds later, you know, you and I, Ariel, would probably not be having this conversation. Instead, he, mostly on gut instinct, came to the conclusion that there was something wrong and said, “This is a false alarm.” And we’re incredibly grateful for that level-headed action of him. He passed away recently.</p><p>His two children are living on very modest means outside of Moscow and we felt that when someone does something like this, or in his case abstains from doing something, that future generations really appreciate, we should show our appreciation, so that others in his situation later on know that if they sacrifice themselves for the greater good, they will be appreciated. Or if they’re dead, their loved ones will. So we organized a ceremony in New York City and invited them to it and bought air tickets for them and so on. And in a very darkly humorous illustration of how screwed up their relationships are at the global level now, the US decided that because — that the way to show appreciation for the US not having gotten nuked was to deny a visa to Stanislav’s son. So he could only join by Skype. Fortunately, his daughter was able to get a visa, even though the waiting period to even get a visa point for Moscow was 300 days. We had to fly her to Israel to get her the Visa.</p><p>But she came and it was her first time ever outside of Russia. She was super excited to come and see New York. It was very touching for me to see all the affection that the New Yorkers there deemed at her and see her reaction and her husband’s reaction and to get to give her this $50,000 award, which for them was actually a big deal. Although it’s of course nothing compared to the value for the rest of the world of what their father did. And it was a very sobering reminder that we’ve had dozens of near misses where we almost had a nuclear war by mistake. And even though the newspapers usually make us worry about North Korea and Iran, of course by far the most likely way in which we might get killed by a nuclear explosion is because another just stupid malfunction or error causing the US and Russia to start a war by mistake.</p><p>I hope that this ceremony and the one we did the year before also, for family of Vasili Arkhipov, can also help to remind people that hey, you know, what we’re doing here, having 14,000 hydrogen bombs and just relying on luck year after year isn’t a sustainable long-term strategy and we should get our act together and reduce nuclear arsenals down to the level needed for deterrence and focus our money on more productive things.</p><p><strong>Ariel:</strong> So I wanted to just add a quick follow-up to that because I had the privilege of attending the ceremony and I got to meet the Petrovs. And one of the things that I found most touching about meeting them was their own reaction to New York, which was in part just an awe of the freedom that they felt. And I think, especially, this is sort of a US centric version of hope, but it’s easy for us to get distracted by how bad things are because of what we see in the news, but it was a really nice reminder of how good things are too.</p><p><strong>Max:</strong> Yeah. It’s very helpful to see things through other people’s eyes and in many cases, it’s a reminder of how much we have to lose if we screw up.</p><p><strong>Ariel:</strong> Yeah.</p><p><strong>Max:</strong> And how much we have that we should be really grateful for and cherish and preserve. It’s even more striking if you just look at the whole planet, you know, in a broader perspective. It’s a fantastic, fantastic place, this planet. There’s nothing else in the solar system even remotely this nice. So I think we have a lot to win if we can take good care of it and not ruin it. And obviously, the quickest way to ruin it would be to have an accidental nuclear war, which — it would be just by far the most ridiculously pathetic thing humans have ever done, and yet, this isn’t even really a major election issue. Most people don’t think about it. Most people don’t talk about it. This is, of course, the reason that we, with the Future of Life Institute, try to keep focusing on the importance of positive uses of technology, whether it be nuclear technology, AI technology, or biotechnology, because if we use it wisely, we can create such an awesome future, like you said: Take the good things we have, make them even better.</p><p><strong>Ariel:</strong> So this seems like a good moment to introduce another guest, who just did a whole podcast series exploring existential risks relating to AI, biotech, nanotech, and all of the other technologies that could either destroy society or help us achieve incredible advances if we use them right.</p><p><strong>Josh:</strong> I’m Josh Clark. I’m a podcaster. And I’m the host of a podcast series called the End of the World with Josh Clark.</p><p><strong>Ariel:</strong> All right. I am really excited to have you on the show today because I listened to all of the End of the World. And it was great. It was a really, really wonderful introduction to existential risks.</p><p><strong>Josh:</strong> Thank you.</p><p><strong>Ariel:</strong> I highly recommend it to anyone who hasn’t listened to it. But now that you’ve just done this whole series about how things can go horribly wrong, I thought it would be fun to bring you on and talk about what you’re still hopeful for after having just done that whole series.</p><p><strong>Josh:</strong> Yeah, I’d love that, because a lot of people are hesitant to listen to the series because they’re like, well, “it’s got to be such a downer.” And I mean, it is heavy and it is kind of a downer, but there’s also a lot of hope that just kind of emerged naturally from the series just researching this stuff. There is a lot of hope — it’s pretty cool.</p><p><strong>Ariel:</strong> That’s good. That’s exactly what I want to hear. What prompted you to do that series, The End of the World?</p><p><strong>Josh:</strong> Originally, it was just intellectual curiosity. I ran across a Bostrom paper in like 2005 or 6, my first one, and just immediately became enamored with the stuff he was talking about — it’s just baldly interesting. Like anyone who hears about this stuff can’t help but be interested in it. And so originally, the point of the podcast was, “Hey, everybody come check this out. Isn’t this interesting? There’s like, people actually thinking about this kind of stuff and talking about it.” And then as I started to interview some of the guys at the Future of Humanity Institute, started to read more and more papers and research further, I realized, wait, this isn’t just like, intellectually interesting. This is real stuff. We’re actually in real danger here.</p><p>And so as I was creating the series, I underwent this transition for how I saw existential risks, and then ultimately how I saw humanity’s future, how I saw humanity, other people, and I kind of came to love the world a lot more than I did before. Not like I disliked the world or people or anything like that. But I really love people way more than I did before I started out, just because I see that we’re kind of close to the edge here. And so the point of why I made the series kind of underwent this transition, and you can kind of tell in the series itself where it’s like information, information, information. And then now, that you have bought into this, here’s how we do something about it.</p><p><strong>Ariel:</strong> So you have two episodes that go into biotechnology and artificial intelligence, which are two — especially artificial intelligence — they’re both areas that we work on at FLI. And in them, what I thought was nice is that you do get into some of the reasons why we’re still pursuing these technologies, even though we do see these existential risks around them. And so, I was curious, as you were doing your research into the series, what did you learn about, where you were like, “Wow, that’s amazing, that I’m so psyched that we’re doing this, even though there are these risks.”</p><p><strong>Josh:</strong> Basically everything I learned about. I had to learn particle physics to explain what’s going on in large Hadron Collider. I had to learn a lot about AI. I realized when I came into it, that my grasp of AI was beyond elementary. And it’s not like I could actually put together a AGI myself from scratch or anything like that now, but I definitely know a lot more than I did before. With biotech in particular, there was a lot that I learned that I found particularly jarring with the number of accidents that are reported every year, and then more than that, the fact that not every lab in the world has to report accidents. I found that extraordinarily unsettling.</p><p>So kind of from start to finish, I learned a lot more than I knew going into it, which is actually one of the main reasons why it took me well over a year to make the series because I would start to research something and then I’d realized I need to understand the fundamentals of this. So I’d go understand, I’d go learn that, and then there’d be something else I had to learn first, before I could learn something the next level up. So I kept having to kind of regressively research and I ended up learning quite a bit of stuff.</p><p>But I think to answer your question, the thing that struck me the most was learning about physics, about particle physics, and how tenuous our understanding of our existence is, but just how much we’ve learned so far in just the last like century or so, when we really dove into quantum physics, particle physics and just what we know about things. One of the things that just knocked my socks off was the idea that there’s no such thing as particles — like particles, as we think of them are just basically like shorthand. But the rest of the world outside of particle physics has said like, “Okay, particles, there’s like protons and neutrons and all that stuff. There’s electrons. And we understand that they kind of all fit into this model, like a solar system. And that’s how atoms work.”</p><p>That is not at all how atoms work, like a particle is just a pack of energetic vibrations and everything that we experience and see and feel, and everything that goes on in the universe is just the interaction of these energetic vibrations in force fields that are everywhere at every point in space and time. And just to understand that, like on a really fundamental level, changed my life actually, changed the way that I see the universe and myself and everything actually.</p><p><strong>Ariel:</strong> I don’t even know where I want to go next with that. I’m going to come back to that because I actually think it connects really nicely to the idea of existential hope. But first I want to ask you a little bit more about this idea of getting people involved more. I mean, I’m coming at this from something of a bubble at this point where I am surrounded by people who are very familiar with the existential risks of artificial intelligence and biotechnology. But like you said, once you start looking at artificial intelligence, if you haven’t been doing it already, you suddenly realize that there’s a lot there that you don’t know.</p><p><strong>Josh:</strong> Yeah.</p><p><strong>Ariel:</strong> I guess I’m curious, now that you’ve done that, to what extent do you think everyone needs to? To what extent do you think that’s possible? Do you have ideas for how we can help people understand this more?</p><p><strong>Josh:</strong> Yeah you know, that really kind of ties into taking on existential risks in general, is just being an interested curious person who dives into the subject and learns as much as you can, but that at this moment in time, as I’m sure you know, that’s easier said than done. Like you really have to dedicate a significant portion of your life to spending time focusing on that one issue whether it’s AI, it’s biotech or particle physics, or nanotech, whatever. You really have to immerse yourself into it because it’s not a general topic of national or global conversation, the existential risks that we’re facing, and certainly not the existential risks we’re facing from all the technology that everybody’s super happy that we’re coming out with.</p><p>And I think that one of the first steps to actually taking on existential risks is for more and more people to start talking about it. Groups like yours, talking to the public, educating the public. I’m hoping that my series did something like that, just arousing curiosity in people, but also raising awareness of these things like, these are real things, these aren’t crackpots talking about this stuff. This is real, legitimate issues that are coming down the pike, that are being pointed out by real, legitimate scientists and philosophers and people who have given great thought about this. This isn’t like a chicken little situation; This is quite real. I think if you can pique someone’s curiosity just enough that they listen, stop and listen, do a little research, it sinks in after a minute that this is real. And that, oh, this is something that they want to be a part of doing something about.</p><p>And so I think just getting people talking about that just by proxy will interest other people who hear about it, and it will spread further and further out. And I think that that’s step one, is to just make it so it’s an okay thing to talk about, so you’re not nuts to raise this kind of stuff seriously.</p><p><strong>Ariel:</strong> Well, I definitely appreciate you doing your series for that reason. I’m hopeful that that will help a lot.</p><p><strong>Ariel:</strong> Now, Allison — you’ve got this website which, my understanding is that you’re trying to get more people involved in this idea that if we focus on these better ideals for the future, we stand a better shot at actually hitting them.</p><p><strong>Allison:</strong> At ExistentialHope.com, I keep a map of reading, podcasts, organizations, and people that inspire an optimistic long-term vision for the future.</p><p><strong>Ariel:</strong> You’re clearly doing a lot to try to get more people involved. What is it that you’re trying to do now, and what do you think we all need to be doing more of to get more people thinking this way?</p><p><strong>Allison:</strong> I do think that it’s up to everyone, really, to try to, again, engage with the fact that we may not be doomed, and what may be on the other side. What I’m trying to do with the website, at least, is generating common knowledge to catalyze more directed coordination toward beautiful futures. I think that there’s a lot of projects out there that are really dedicated to identifying the threats to human existence, but very few really offer guidance on what to influence that. So I think we should try to map the space of both peril and promise which lie before us, but we should really try to aim for that this knowledge can empower each and every one of us to navigate toward the grand future.</p><p>For us currently on the website this involves orienting ourselves, so collecting useful models, and relevant broadcasts, and organizations that generate new insights, and then try to synthesize a map of where we came from, and a really kind of long perspective, and where we may go, and then which lenses of science and technology and culture are crucial to consider along the way. Then finally we would like to publish a living document that summarizes those models that are published elsewhere, to outline possible futures, and the idea is that this is a collaborative document. Even already, currently, the website links to a host of different Google docs in which we’re trying to really synthesize the current state of the art in the different focus areas. The idea is that this is collaborative. This is why it’s on Google docs, because everyone can just comment. And people do, and I think this should really be a collaborative effort.</p><p><strong>Ariel: </strong>What are some of your favorite examples of content that, presumably, you’ve added to your website, that look at these issues?</p><p><strong>Allison:</strong> There’s quite a host of things on there, I think, that a good start for people to go on the website is just to go on the overview. Because here I list kind of my top 10 lists about short pieces and long pieces, but my personal ones, I think, as a starting ground: I really like the metaethics sequence by Eliezer Yudkowsky. It contains a really good post, like Existential Angst Factory, and Reality as Fixed Computation. For me this is kind of like existentialism 2.0. Have to get your motivations and expectations right. What can I reasonably hope for? Then I think, relatedly, there’s also the Fan Sequence, also by Yudkowsky. But that together with, for example, Letter From Utopia by Nick Bostrom, or Hedonistic Imperative by David Pearce, or Post On Raikoth by Scott Alexander — they are really a nice next step because they actually lay out a few compelling positive versions of utopia.</p><p>Then if you want to get into the more nitty gritty there’s a longer section on civilization, its past and its future — so, what’s wrong and how to improve it. Here Nick Bostrom wrote this piece on the future of human evolution, which lays out two suboptimal paths for humanity’s future, and interestingly enough they don’t involve extinction. A similar one, I think, which probably many people are familiar with, is Scott Alexander’s Meditations On Moloch, and then some that people are less familiar with — Growing Children For Bostrom’s Disneyland. They are really interesting, because they are other pieces of this type, which are sketching out competitive and selective pressures that lead toward races to the bottom, as negative futures which don’t involve extinction per se. I think the really interesting thing, then, is that even those features are only bad if you think that the bottom is bad.</p><p>Next to them I list books, for example, like Robin Hanson, Age of M, which argues that living at subsistence may not be terrible, and in fact it’s pretty much what most of our past lives outside of the current dream time have always involved. So I think those are two really different lenses to make sense of the same reality, and I personally found this contrast so intriguing that I hosted a salon last year with Paul Christiano, Robin Hanson, Peter Eckersley, and a few others to kind of map out where we may be racing towards, so how bad those competitive equilibria actually are. I also link to those from the website.</p><p>To me it’s always interesting to map out one potentially possible future visions, and then try to find one either that contradicts or compliments it. I think having a good idea of an overview of those gives you a good map, or at least a space of possibilities.</p><p><strong>Ariel:</strong> What do you recommend to people who are interested in trying to do more? How do you suggest they get involved?</p><p><strong>Allison:</strong> One thing, an obvious thing, would be commenting on the Google Docs, and I really encourage everyone to do that. Another one would be just to join the mailing list. You can kind of indicate whether you want updates on me, or whether you want to collaborate, in which case we may be able to reach out to you. Or if you’re interested in meetups, they would only be in San Francisco so far, but I’m hoping that there may be others. I do think that currently the project is really in its infancy. We are relying on the community to help with this, so there should be a kind of collaborative vision.</p><p>I think that one of the main things that I’m hoping that people can get out of it for now is just to give some inspiration on where we may end up if we get it right, and on why work toward better futures, or even work toward preventing existential risks, is both possible and necessary. If you go on the website on the first section — the vision section — that’s what that section is for.</p><p>Secondly, then, if you are already opted in, if you’re already committed, I’m hoping that perhaps the project can provide some orientation. If someone would like to help but doesn’t really know where to start, the focus areas are an attempt to map out the different areas that we need to make progress on for better futures. Each area comes with an introductory text, and organizations that are working in that area that one can join or support, and Future of Life is in a lot of those areas.</p><p>Then I think finally, just apart from inspiration or orientation, it’s really a place for collaboration. The project is in its infancy and everyone should contribute their favorite pieces to our better futures.</p><p><strong>Ariel:</strong> I’m really excited to see what develops in the coming year for existentialhope.com. And, naturally, I also want to hear from Max and Anthony about 2019. What are you looking forward to for FLI next year?</p><p><strong>Max:</strong> For 2019 I’m looking forward to more constructive collaboration on many aspects of this quest for a good future for everyone on earth. At the nerdy level, I’m looking forward to more collaboration on AI’s safety research and also ways of making the economy, that keeps growing thanks to AI, actually make everybody better off, rather than some people poorer and angrier. And at the most global level really looking forward to working harder to get past this outdated us versus them attitude that we still have between the US and China and Russia and other major powers. Many of our political leaders are so focused on the zero sum game mentality that they will happily risk major risks of nuclear war and AI arms races and other outcomes where everybody would lose, instead of just realizing hey, you know, we’re actually in this together. What does it mean for America to win? It means that all Americans get better off. What does it mean for China to win? It means that the Chinese people all get better off. Those two things can obviously happen at the same time as long as there’s peace, and technology just keeps improving life for everybody.</p><p>In practice, I’m very eagerly looking forward to seeing if we can get scientists from around the world — for example, AI researchers — to converge on certain shared goals that are really supported everywhere in the world, including by political leaders and in China and the US and Russia and Europe and so on, instead of just obsessing about the differences. Instead of thinking us versus them, it’s all of us on this planet working together against the common enemy, which is our own stupidity and the tendency to make bad mistakes, so that we can harness this powerful technology to create a future where everybody wins.</p><p><strong>Anthony:</strong> I would say I’m looking forward to more of what we’re doing now, thinking more about the futures that we do want. What exactly do those look like? Can we really think through pictures of the future that makes sense to us that are attractive, that are plausible, and yet aspirational, and where we can identify things and systems and institutions that we can build now toward the aim of getting us to those futures? I think there’s been a lot of, so far, thinking about what are the major problems that might arise, and I think that’s really, really important, and that project is certainly not over, and it’s not like we’ve avoided all of those pitfalls by any means, but I think it’s important not to just not fall into the pit, but to actually have a destination that we’d like to get to — you know, the resort at the other end of the jungle or whatever.</p><p>I find it frustrating a bit when people do what I’m doing now: they talk about talking about what we should and shouldn’t do. But they don’t actually talk about what we should and shouldn’t do. I think the time has come to actually talk about it in the same way that when… there was the first use of CRISPR in a embryo that came to term. So everybody’s talking about, “Well, we need to talk about what we should and shouldn’t do with this. We need to talk about that, we need to talk about it.” Let’s talk about it already.</p><p>So I’m excited about upcoming events that FLI will be involved in that are explicitly thinking about: let’s talk about what that future is that we would like to have and let’s debate it, let’s have that discussion about what we do want and don’t want, try to convince each other and persuade each other of different visions for the future. I do think we’re starting to actually build those visions for what institutions and structures in the future might look like. And if we have that vision, then we can think of what are the things we need to put in place to have that.</p><p><strong>Ariel:</strong> So one of the reasons that I wanted to bring Gaia on is because I’m working on a project with her — and it’s her project — where we’re looking at this process of what’s known as worldbuilding, to sort of look at how we can move towards a better future for all. I was hoping you could describe it, this worldbuilding project that I’m attempting to help you with, or work on with you. What is worldbuilding, and how are you modifying it for your own needs?</p><p><strong>Gaia:</strong> Yeah. Worldbuilding is a really fascinating set of techniques. It’s a process that has its roots in narrative fiction. You can think of, for example, the entire complex world that J.R.R. Tolkien created for The Lord of the Rings series, for example. And in more contemporary times, some spectacularly advanced worldbuilding is occurring now in the gaming industry. So these huge connected systems of systems that underpin worlds in which millions of people today are playing, socializing, buying and selling goods, engaging in an economy. These are these vast online worlds that are not just contained on paper as in a book, but are actually embodied in software. And over the last decade, world builders have begun to formally bring these tools outside of the entertainment business, outside of narrative fiction and gaming, film and so on, and really into society and communities. So I really define worldbuilding as a powerful act of creation.</p><p>And one of the reasons that it is so powerful is that it really facilitates collaborative creation. It’s a collaborative design practice. And in my personal definition of worldbuilding, the way that I’m thinking of it, and using it, is that it unfolds in four main stages. The first stage is: we develop a foundation of shared knowledge that’s grounded in science, and research, and relevant domain expertise. And the second phase is building on that foundation of knowledge. We engage in an exercise where we predict how the interconnected systems that have emerged in this knowledge database — we predict how they will evolve. And we imagine the state of their evolution at a specific point in the future. Then the third phase is really about capturing that state in all its complexity, and making that information useful to the people who need to interface with it. And that can be in the form of interlinked databases and particularly also in the form of visualizations, which help make these sort of abstract ideas feel more present and concrete. And then the fourth and final phase is then utilizing that resulting world as a tool that can be used to support scenario simulation, research, and development in many different areas including public policy, media production, education, and product development.</p><p>I mentioned that these techniques are being brought outside of the realm of entertainment. So rather than just designing fantasy worlds for the sole purpose of containing narrative fiction and stories, these techniques are now being used with communities, and Fortune 500 companies, and foundations, and NGOs, and other places, to create plausible future worlds. It’s fascinating to me to see how these are being used. For example, they’re being used to reimagine the mission of an organization. They’re being used to plan for the future, and plan around a collective vision of that future. They’re very powerful for developing new strategies, new programs, and new products. And I think to me one of the most interesting things is really around informing policy work. That’s how I see worldbuilding.</p><p><strong>Ariel:</strong> Are there any actual examples that you can give or are they proprietary?</p><p><strong>Gaia:</strong> There are many examples that have created some really incredible outcomes. One of the first examples of worldbuilding that I ever learned about was a project that was done with a native Alaskan tribe. And the comments that came from the tribe and about that experience were what really piqued my interest. Because they said things like, “This enabled us to sort of leap frog over the barriers in our current thinking and imagine possibilities that were sort of beyond what we had considered.” This project brought together several dozen members of the community, again, to engage in this collaborative design exercise, and actually visualize and build out those systems and understand how they would be interconnected. And it ended up resulting in, I think, some really incredible things. Like a partnership with MIT where they brought a digital fabrication lab onto their reservation, and created new education programs around digital design and digital fabrication for their youth. And there’s a lot of other things that are still coming out of that particular worldbuild.</p><p>There are other examples where Fortune 500 companies are building out really detailed, long-term worldbuilds that are helping them stay relevant, and imagine how their business model is going to need to transform in order to adapt to really plausible, probable futures that are just around the corner.</p><p><strong>Ariel:</strong> I want to switch now to what you specifically are working on. The project we’re looking at is looking roughly 20 years into the future. And you’ve sort of started walking through a couple systems yourself while we’ve been working on the project. And I thought that it might be helpful if you could sort of walk through, with us, what those steps are to help understand how this process works.</p><p><strong>Gaia:</strong> Maybe I’ll just take a quick step back, if that’s okay and just explain the worldbuild that we’re preparing for.</p><p><strong>Ariel:</strong> Yeah. Please do.</p><p><strong>Gaia:</strong> This is a project called Augmented Intelligence. The first Augmented Intelligence summit is happening in March in 2019. And our goal with this project is really to engage with and shift the culture, and also our mindset, about the future of artificial intelligence. And to bring together a multidisciplinary group of leaders from government, academia, and industry, and to do a worldbuild that’s focused on this idea of: what does our future world look like with advanced AI deeply integrated into it? And to go through the process of really imagining and predicting that world in a way that’s just a bit further beyond the horizon that we normally see and talk about. And that exercise, that’s really where we’re getting that training for long-term thinking, and for systems level thinking. And the world that results — our hope is that it will allow us to develop better intuitions, to experiment, to simulate scenarios, and really to have a more attuned capacity to engage in many ways with this future. And ultimately explore how we want to evolve our tools and our society to meet that challenge.</p><p><strong>Gaia:</strong> What will come out of this process — it really is a generative process that will create assets and systems that are interconnected, that inhabit and embody a world. And this world should allow us to experiment, and simulate scenarios, and develop a more attuned capacity to engage with the future. And that means on both an intuitive level and also in a more formal structured way. And ultimately our goal is to use this tool to explore how we want to evolve as a society, as a community, and to allow ideas to emerge about what solutions and tools will be needed to adapt to that future. Our goal is to really bootstrap a steering mechanism that allows us to navigate more effectively toward outcomes that support human flourishing.</p><p><strong>Ariel:</strong> I think that’s really helpful. I think an example to walk us through what that looks like would be helpful.</p><p><strong>Gaia:</strong> Sure. You know, basically what would happen in a worldbuilding process is that you would have some constraints or some sort of seed information that you think is very likely — based on research, based on the literature, based on sort of the input that you’re getting from domain experts in that area. For example, you might say, “In the future we think that education is all going to happen in a virtual reality system that’s going to cover the planet.” Which I don’t think is actually the case, but just to give an example. You might say something like, “If this were true, then what are the implications of that?” And you would build a set of systems, because it’s very difficult to look at just one thing in isolation.</p><p>Because as soon as you start to do that — John Muir says, “As soon as you try to look at just one thing, you find that it is irreversibly connected to everything else in the universe.” And I apologize to John Muir for not getting that quote exactly correct, he says it much more eloquently than that. But the idea is there. And that’s sort of what we leverage in a worldbuilding process: where you take one idea and then you start to unravel all of the implications, and all of the interconnecting systems that would be logical, and also possible, if that thing were true. It really does depend on the quality of the inputs. And that’s something that we’re working really, really hard to make sure that our inputs are believable and plausible, but don’t put too much in terms of constraints on the process that unfolds. Because we really want to tap into the creativity in the minds of this incredible group of people that we’re gathering, and that is where the magic will happen.</p><p><strong>Ariel:</strong> To make sure that I’m understanding this right: if we use your example of, let’s say all education was being taught virtually, I guess questions that you might ask or you might want to consider would be things like: who teaches it, who’s creating it, how do students ask questions, who would their questions be directed to? What other types of questions would crop up that we’d want to consider? Or what other considerations do you think would crop up?</p><p><strong>Gaia:</strong> You also want to look at the infrastructure questions, right? So if that’s really something that is true all over the world, what do server farms look like in that future, and what’s the impact on the environment? Is there some complimentary innovation that has happened in the field of computing that has made computing far more efficient? How have we been able to do this? Given the — there are certain physical limitations that just exist on our planet. If X is true in this interconnected system, then how have we shaped, and molded, and adapted everything around it to make that thing true? You can look at infrastructure, you can look at culture, you can look at behavior, you can look at, as you were saying, communication and representation in that system and who is communicating. What are the rules? I mean, I think a lot about the legal framework, and the political structure that exists around this. So who has power and agency? How are decisions made?</p><p><strong>Ariel:</strong> I don’t know what this says about me, but I was just wondering what detention looks like in a virtual world.</p><p><strong>Gaia:</strong> Yeah. It’s a good question. I mean, what are the incentives and what are the punishments in that society? And do our ideas of what incentives and punishments look like actually change in that context? There isn’t a place where you can come on a Saturday if there’s no physical school yard. How is detention even enforced when people can log in and out of the system at will?</p><p><strong>Ariel:</strong> All right, now you have me wondering what recess looks like.</p><p><strong>Gaia:</strong> So you can see that there are many different fascinating sort of rabbit holes that you could go down. And of course our goal is to really make this process really useful to imagining the way that we want our policies, and our tools, and our education to evolve.</p><p><strong>Ariel:</strong> I want to ask one more question about … Well, it’s sort of about this but there’s also a broader aspect to it. And that is, I hear a lot of talk — and I’m one of the people saying this because I think it’s absolutely true — that we need to broaden the conversation and get more diverse voices into this discussion about what we want our future to look like. But what I’m finding is that this sounds really nice in theory, but it’s incredibly hard to actually do in practice. I’m under the impression that that is some of what you’re trying to address with this project. I’m wondering if you can talk a little bit about how you envision trying to get more people involved in considering how we want our world to look in the future.</p><p><strong>Gaia:</strong> Yeah, that’s a really important question. One of the sources of inspiration for me on this point was a conversation with Stuart Russell — an interview with Stuart Russell, I should say — that I listened to. We’ve been really fortunate and we are thrilled that he’s one of our speakers and he’ll be involved in the worldbuilding process. And he kind of talks about this idea that the artificial intelligence researchers, the roboticists, even a few technologists that are building these amplifying tools that are just increasing in potency year over year, are not the only ones who need to have input into the conversation around how they’re utilized and the implications on all of us. And that’s really one of the sort of core philosophies behind this particular project, is that we really want it to be a multidisciplinary group that comes together, and we’re already seeing that. We have a really wonderful set of collaborators who are thinking about ethics in this space, and who are thinking about a broader definition of ethics, and different cultural perspectives on ethics. And how we can create a conversation that allows space for those to simultaneously coexist.</p><p><strong>Allison:</strong> I recently had a similar kind of question that arose in conversation, which was about: why are we lacking positive future visions so much? Why are we all kind of stuck in a snapshot of the current suboptimal macro situation? I do think it’s our inability to really think in larger terms. If you look at our individual human life, clearly for most of us, it’s pretty incredible — our ability to lead much longer and healthier lives than ever before. If we compare this to how well humans used to live, this difference is really unfathomable. I think Yuval Harari said it right, he said “You wouldn’t want to have lived 100 years ago.” I think that’s correct. On the other hand I also think that we’re not there yet.</p><p>I find it, for example, pretty peculiar that we say that we value freedom of choice in everything we do, but in the one thing that’s kind of the basis of all of our freedoms, which is our very existence, we leave it up again to slowly deteriorate according to aging. This would really deteriorate ourselves and everything we value. I think that every day aging is burning libraries. We’ve come a long way, but we’re not safe, and we are definitely not there yet. I think the same holds true for civilization at large. I think thanks to a lot of technologies our living standards have been getting better and better, and I think the decline of poverty and violence are just a few examples.</p><p>We can share knowledge much easier, and I think everyone who’s read Enlightenment Now will be kind of tired of those graphs, but again, I also think that we’re not there yet. I think even though we have less wars than ever before, the ability to wipe ourselves out as a species also really exists, and I think in fact this ability is now more available to more people, and with technologies of maturity, it may really only take a small and well-curated group of individuals to cause havoc of catastrophic consequences. If you let that sink in, it’s really absurd that we have no emergency plan for the use of technological weapons. We have no plans to rebuild civilization. We have no plans to back up human life.</p><p>I think that current news articles take too much of a short term view. They’re more a snapshot. I think the long-term view, on the one hand, opens up this eye of, “Hey, look how far we’ve come,” but also, “Oh man. We’re here, and we’ve made it so far, but there’s no feasible plan for safety yet.” I do think we need to change that, so I think the long run doesn’t only open up rosy glasses, but also the realization that we ought to do more because we’ve come so far.</p><p><strong>Josh:</strong> Yeah, one of the things that makes this time so dangerous is we’re at this kind of a fork in the road, where if we go this one way, like say, with figuring out how to develop friendliness in AI, we could have this amazing, astounding future for humanity that stretches for billions and billions and billions of years. One of the things that really opened my eyes was, I always thought that the heat death of the universe will spell the end of humanity. There’s no way we’ll ever make it past that, because that’s just the cessation of everything that makes life happen, right? And we will probably have perished long before that. But let’s say we figured out a way to just make it to the last second and humanity dies at the same time the universe does. There’s still an expiration date on humanity. We still go extinct eventually. But one of the things I ran across when I was doing research for the physics episode is that the concept of growing a universe from seed, basically, in a lab is out there. It’s done. I don’t remember who came up with it. But somebody has sketched out basically how to do this.</p><p>It’s 2018. If we think 100 or 200 or 500 or a thousand years down the road and that concept can be built upon and explored, we may very well be able to grow universes from seed in laboratories. Well, when our universe starts to wind down or something goes wrong with it, or we just want to get away, we could conceivably move to another universe. And so we suddenly lose that expiration date for humanity that’s associated with the heat death of the universe, if that is how the universe goes down. And so this idea that we have a future lifetime that spans into at least the multiple billions of years — at least a billion years if we just manage to stay alive on Planet Earth and never spread out but just don’t actually kill ourselves — when you take that into account the stakes become so much higher for what we’re doing today.</p><p><strong>Ariel:</strong> So, we’re pretty deep into this podcast, and we haven’t heard anything from Anders Sandberg yet, and this idea that Josh brought up ties in with his work. Since we’re starting to talk about imagining future technologies, let’s meet Anders.</p><p><strong>Anders:</strong> Well, I’m delighted to be on this. I’m Anders Sandberg. I’m a senior research fellow at The Future of Humanity Institute at University of Oxford.</p><p><strong>Ariel:</strong> One of the things that I love, just looking at your FHI page, you talk about how you try to estimate the capabilities of future technology. I was hoping you could talk a little bit about what that means, what you’ve learned so far, how one even goes about studying the capabilities of future technologies?</p><p><strong>Anders:</strong> Yeah. It is a really interesting problem because technology is based on ideas. As a general rule, you cannot predict what ideas people will come up with in the future, because if you could, you would already kind of have that idea. So this means that, especially technologies that are strongly dependent on good ideas, are going to be tremendously hard to predict. This is of course why artificial intelligence is a little bit of a nightmare. Similarly, biotechnology is strongly dependent on what we discover in biology and a lot of that is tremendously weird, so again, it’s very unpredictable.</p><p>Meanwhile, other domains of life are advancing at a more sedate pace. It’s more like you incrementally improve things. So the ideas are certainly needed, but we don’t really change everything around. If you think about more slower, microprocessors are getting better and a lot of improvements are small, incremental ones. Some of them require a lot of intelligence to come up with, but in the end it all sums together. It’s a lot of small things adding together. So you can see a relatively smooth development in the large.</p><p><strong>Ariel: </strong>Okay. So what you’re saying is we don’t just have each year some major discovery, and that’s what doubles it. It’s lots of little incremental steps.</p><p><strong>Anders:</strong> Exactly. But if you look at the performance of some software, quite often it goes up smoothly because the computers are getting better and then somebody has a brilliant idea that can do it not just in 10% less time, but maybe in 10% of the time that it would have taken. For example, the fast Fourier transform that people invented in the 60s and 70s enables the compression we use today for video and audio and enables multimedia on the internet. Without that to speed up, it would not be practical to do, even with current computers. This is true for a lot of things in computing. You get a surprise insight and the problem that previously might be impossible to do efficiently suddenly becomes quite convenient. So the problem is of course: what can we say about the abilities of future technology if these things happen?</p><p>One of the nice things you can do is you can lean on the laws of physics. There are good reasons not to think that perpetual motion machines can work, because we understand, actually, energy conservation and the laws of thermodynamics that give very strong reason why this cannot happen. We can be pretty certain that that’s not possible. We can analyze what would then be possible if you had perpetual motion machines or faster than light transport and you can see that some of the consequences are really weird. But it makes you suspect that this is probably not going to happen. So that’s one way of looking at it. But you can do the reverse: You can take laws of physics and engineering that you understand really well and make fictional machines — essentially work out all the details and say “okay, I can’t build this but were I to build it, in that case what properties would it have?” If I wanted to build, let’s say, a machine made out of atoms, could I make it to work? And it turns out that this is possible to do in a rigorous way, and it tells you capabilities about machines that don’t exist yet, and maybe we will never build, but it shows you what’s possible.</p><p>This is what Eric Drexler did for nanotechnology in the 80s and 90s. He basically worked out what would be possible if we could put atoms in the right place. He could demonstrate that this would produce machines of tremendous capability. We still haven’t built them, but he proved that these can be built — and we probably should build them because they are so effective, so environmentally friendly, and so on.</p><p><strong>Ariel:</strong> So you gave the example of what he came up with a while back. What sort of capabilities have you come across that you thought were interesting that you’re looking forward to us someday pursuing?</p><p><strong>Anders:</strong> I’ve been working a little bit on the questions about “is it possible to settle a large part of the universe?” I have been working out, together with my colleagues, a bit of the physical limitations of that. All in all, we found that a civilization doesn’t need to use an enormous, astronomical amount of matter and energy to settle a very large chunk of the universe. The total amount of matter corresponds with roughly a Mercury-sized planet in a solar system in each of the galaxies. Many people would say if you want to settle the universe you need an enormous spacecraft and you need enormous amount of energy. It looks like you would be able to see that across half of the universe, but we could demonstrate that actually if you essentially use matter from a really big asteroid or a small planet, you can get enough solar collectors to launch small spacecraft to all the stars and all the galaxies within reach and there you’ll use again a bit of asteroids to do it. The laws of physics allow intelligent life to spread across an enormous amount of the universe in a rather quiet way.</p><p><strong>Ariel: </strong>So does that mean you think it’s possible that there is life out there and it’s reasonable for us not to have found it?</p><p><strong>Anders:</strong> Yes. If we were looking at the stars, we would probably miss if one or two stars in remote galaxies were covered with solar collectors. It’s rather easy to miss them among the hundreds of billions of other stars. This was actually the reason we did this paper: We demonstrate that much of the thinking about the Fermi paradox — that annoying question that well, there ought to be a lot of intelligent life out in the universe given how large it is and that we tend to think that it’s relatively likely yet we don’t see anything — many of those explanations are based on the possibility of colonizing just the Milky Way. In this paper, we demonstrate that actually you need to care about all the other galaxies too. In a sense, we made the fermi paradox between a million and a billion times worse. Of course, this is all in a day’s work for us in the Philosophy Department, making everybody’s headaches bigger.</p><p><strong>Ariel:</strong> And now it’s just up to someone else to figure out the actual way to do this technically.</p><p><strong>Anders:</strong> Yeah, because it might actually be a good idea for us to do.</p><p><strong>Ariel:</strong> So Josh, you’ve mentioned the future of humanity a couple of times, and humanity in the future, and now Anders has mentioned the possibility of colonizing space. I’m curious how you think that might impact humanity. How do you define humanity in the future?</p><p><strong>Josh:</strong> I don’t know. That’s a great question. It could take any number of different routes. I think — Robin Hanson is an economist who came up with this, the great filter hypothesis, and I talked to him about that very question. His idea was that — and I’m sure it’s not just his, but it’s probably a pretty popular idea — that once we spread out from Earth and start colonizing further and further out into the galaxy, and then into the universe, we’ll undergo speciation events like, there will be multiple species of humans in the universe again, just like there was like 50,000 years ago, when we shared Earth with multiple species of humans.</p><p>The same thing is going to happen as we spread out from Earth. I mean, I guess the question is, which humans are you talking about, in what galaxy? I also think there’s a really good chance — and this could happen among multiple human species — that at least some humans will eventually shed their biological form and upload themselves into some sort of digital format. I think if you just start thinking in efficiencies, that’s just a logical conclusion to life. And then there’s any number of routes we could take and change especially as we merge more with technology or spread out from Earth and separate ourselves from one another. But I think the thing that really kind of struck me as I was learning all this stuff is that we tend to think of ourselves as the pinnacle of evolution, possibly the most intelligent life in the entire universe, right? Certainly the most intelligent on Earth, we’d like to think. But if you step back and look at all the different ways that humans can change, especially like the idea that we might become post-biological, it becomes clear that we’re just a point along a spectrum that keeps on stretching out further and further into the future than it does even into the past.</p><p>We’re just at a current situation on that point right now. We’re certainly not like the end-all be-all of evolution. And ultimately, we may take ourselves out of evolution by becoming post-biological. It’s pretty exciting to think about all the different ways that it can happen, all the different routes we can take — there doesn’t have to just be one single one either.</p><p><strong>Ariel:</strong> Okay, so, I kind of want to go back to some of the space stuff a little bit, and Anders is the perfect person for my questions. I think one of the first things I want to ask is, very broadly, as you’re looking at these different theories about whether or not life might exist out in the universe and that it’s reasonable for us not to have found it, do you connect the possibility that there are other life forms out there with an idea of existential hope for humanity? Or does it cause you concern? Or are they just completely unrelated?</p><p><strong>Anders:</strong> The existence of extraterrestrial intelligence: if we knew they existed that would in some sense be hopeful because we know the universe allows for more than our kind of intelligence and intelligence might survive over long spans of time. If we just discovered that we’re all alone and a lot of ruins from extinct civilizations, that would be very bad news for us. But we might also have this weird situation that we currently feel, that we don’t see anybody. We don’t notice any ruins; Maybe we’re just really unique and should perhaps feel a bit proud or lucky but also responsible for a whole universe. It’s tricky. It seems like we could learn something very important if we understood how much intelligence there is out there. Generally, I have been trying to figure out: is the absence of aliens evidence for something bad? Or might it actually be evidence for something very hopeful?</p><p><strong>Ariel:</strong> Have you concluded anything?</p><p><strong>Anders:</strong> Generally, our conclusion has been that the absence of aliens is not surprising. We tend to think that the Fermi Paradox implies “oh, there’s something strange here.” The universe is so big and if you multiply the number of stars with some reasonable probability, you should get loads of aliens. But actually, the problem here is reasonable probability. We normally tend to think of that as something like bigger than one chance in a million or so, but actually, there is no reason the laws of physics wouldn’t put a probability that’s one in a googol. It actually turns out that we’re uncertain enough about the origin of life and the origins of intelligence and other forms of complexity that it’s not implausible that maybe we are the only life within the visible universe. So we shouldn’t be too surprised about that empty sky.</p><p>One possible reason for the great silence is that life is extremely rare. Another possibility might be that life is not rare but it’s very rare that it becomes the kind of life that evolves to complex nervous systems. Another reason might be of course that once you get intelligence, well, it destroys itself relatively quickly, and Robin Hanson has called this the Great Filter. We know that one of the terms in the big equation for the number of civilizations in the universe needs to be very small; otherwise, the sky would be full of aliens. But is that one of the early terms, like the origin of life, or the origin of intelligence — or the late term, how long intelligence survives? Now, if there is an early Great Filter, this is rather good news for us. We are going to be very unique and maybe a bit lonely, but, it doesn’t tell us anything dangerous about our own chances. Of course, we might still flub it and go extinct because our own stupidity but that’s kind of up to us rather than the laws of physics.</p><p>On the other hand, if it turns out that there is a late Great Filter, then even though we know the universe might be dangerous, we’re still likely to get wiped out — which is very scary. So, figuring out where the unlikely terms in the big equation are is actually quite important for making a guess about our own chances.</p><p><strong>Ariel: </strong>Where are we now in terms of that?</p><p><strong>Anders:</strong> Right now, in my opinion — I have a paper, not published yet but it’s in the review process, where we try to apply proper uncertainty calculations to this. Because many people make guesstimates about the probabilities of various things, admit that they’re guesstimates, and then get a number at the end that we also admit is a bit uncertain. But we haven’t actually done a proper uncertainty calculation so quite a lot of these numbers become surprisingly biased. So instead of saying that maybe there’s one chance in a million that a planet develops life, you should try to have a full range of what’s the lowest probability there could be for life and what’s the highest probability and how do you think it distributes between them. If you use that kind of proper uncertainty range and then multiply it all together and do the maths right, then you get the probability distribution for how many alien species there could be in the universe. Even if you’re starting out as somebody who’s relatively optimistic about the mean value of all of this, you will still find that you get a pretty big chunk of probability that we’re actually pretty alone in the Milky Way or even the observable universe.</p><p>In some sense, this is just common sense. But it’s a very nice thing to be able to quantify the common sense, and then start saying: so what happens if we for example discover that there is life on Mars? What will that tell us? How will that update things? You can use the math to calculate that, and this is what we’ve done. Similarly, if we notice that there doesn’t seem to be any alien super civilizations around the visible universe, that’s a very weak update but you can still use that to see that this updates our estimates of the probability of life and intelligence much more than the longevity of civilizations.</p><p>Mathematically this gives us a reason to think that the Great Filter might be early. The absence of life might be rather good news for us because it means that once you get intelligence, there’s no reason why it can’t persist for a long time and grow into something very flourishing. That is a really good cause of existential hope. It’s really promising, but we of course need to do our observations. We actually need to look for life, we need to look out in the sky and see. You may find alien civilizations. In the end, any amount of mathematics and armchair astrobiology, that’s always going to be disproven by any single observation.</p><p><strong>Ariel: </strong>That comes back to a question that came to mind a bit earlier. As you’re looking at all of this stuff and especially as you’re looking at the capabilities of future technologies, once we figure out what possibly could be done, can you talk a little bit about what our limitations are today from actually doing it? How impossible is it?</p><p><strong>Anders: </strong>Well, impossible is a really tricky word. When I hear somebody say “it’s impossible,” I immediately ask “do you mean against the laws of physics and logic” or “we will not be able to do this for the foreseeable future” or “we can’t do it within the current budget”?</p><p><strong>Ariel: </strong>I think maybe that’s part of my question. I’m guessing a lot of these things probably are physically possible, which is why you’ve considered them, but yeah, what’s the difference between what we’re technically capable of today and what, for whatever reason, we can’t budget into our research?</p><p><strong>Anders: </strong>We have a domain of technologies that we already have been able to construct. Some of them are maybe too expensive to be very useful. Some of them still requires a bunch of grad students holding them up and patching them as they are breaking all the time, but we can kind of build them. And then there’s some technology that we are very robustly good at. We have been making cog wheels and combustion engines for decades now and we’re really good at that. Then there are these technologies that we can do exploratory engineering to demonstrate that if we actually had cog wheels made out of pure diamond or the Dyson shell surrounding the sun collecting energy, they could do the following things.</p><p>So they don’t exist as practical engineering. You can work out blueprints for them and in some sense of course, once we have a complete enough blueprint, if you asked could you build the thing, you could do it. The problem is of course normally you need the tools and resources for that, and you need to make the tools to make the tools, and the tools to make those tools, and so on. So if we wanted to make atomically precise manufacturing today, we can’t jump straight to it. What we need to make is a tool that allows us to build things that are moving us much closer.</p><p>The Wright Brothers’ airplane was really lousy as an airplane but it was flying. It’s a demonstration, but it’s also a tool that allows you to make a slightly better tool. You would want to get through this and you’d probably want to have a roadmap and do experiments and figure out better tools to do that.</p><p>This is typically where scientists actually have to give way to engineers. Because engineers care about solving a problem rather than being the most elegant about it. In science, we want to have this beautiful explanation of how everything works; Then we do experiments to test whether it’s true and refine our explanation. But in the end, the paper that gets published is going to be the one that has the most elegant understanding. In engineering, the thing that actually sells and changes the world is not going to be the most elegant thing but the most useful thing. The AK-47 is in many ways not a very precise piece of engineering but that’s the point. It should be possible to repair it in the field.</p><p>The reason our computers are working so well was we figured out the growth path where you use photolithography to etch silicon chips, and that allowed us to make a lot of them very cheaply. As we learned more and more about how to do that, they became cheaper and more capable and we developed even better ways of etching them. So in order to build molecular nanotechnology, you would need to go through a somewhat similar chain. It might be that you start out with using biology to make proteins, and then you use the proteins to make some kind of soft machinery, and then you use that soft machinery to make hard machinery, and eventually end up with something like the work of Eric Drexler.</p><p><strong>Ariel:</strong> I actually want to step back to the present now and you mentioned computers and we’re doing them very well. But computers are also an example of — or maybe software I suppose is more the example — of technology that works today but it often fails. Especially when we’re considering things like AI safety in the future, what should we make of the fact that we’re not designing software to be more robust? I mean, I think especially if we look at something like airplanes which are quite robust, we can see that it could be done but we’re still choosing not to.</p><p><strong>Anders: </strong>Yeah, nobody would want to fly with an airplane that crashed as often as a word processor.</p><p><strong>Ariel:</strong> Exactly.</p><p><strong>Anders:</strong> It’s true that the earliest airplanes were very crash prone — in fact most of them were probably as bad as our current software is. But the main reason we’re not making software better is that most of the time we’re not willing to pay for that quality. Also, that there is some very hard engineering problems with engineering complexity. So making a very hard material is not easy but in some sense, it’s a straightforward problem. If, on the other hand, you have literally billions of moving pieces that all need to fit together, then it gets tricky to make sure that this always works as it should. But it can be done.</p><p>People have been working on mathematical proofs that certain pieces of software are correct and secure. It’s just that up until recently, it’s been so expensive and tough that nobody really cared to do it except maybe some military groups. Now it’s starting to become more and more essential because we’ve built our entire civilization on a lot of very complex systems that are unfortunately very insecure, very unstable, and so on. Most of the time we get around it by making backup copies and whenever a laptop crashes, well, we reboot it, swear a bit and hopefully we haven’t lost too much work.</p><p>That’s not always a bad solution — a lot of biology is like that too. Cells in our bodies are failing all the time but they’re just getting removed and replaced and then we try again. But this, of course, is not enough for certain sensitive applications. If we ever want to have brain-to-computer interfaces, we certainly want to have good security so we don’t get hacked. If we want to have very powerful AI systems, we want to make sure that their motivations are constrained in such a way that they’re helpful. We also want to make sure that they don’t get hacked or develop weird motivations or behave badly because their owners told them to behave badly. Those are very complex problems: It’s not just like engineering something that’s simply safe. You’re going to need entirely new forms of engineering for that kind of learning system.</p><p>This is something we’re learning. We haven’t been building things like software for very long and when you think about the sheer complexity of a normal operating system, even a small one running on a phone, it’s kind of astonishing that it works at all.</p><p><strong>Allison:</strong> I think that Eliezer Yudkowsky once said that the problem of our complex civilization is its complexity. It does seem that technology is outpacing our ability to make sense of it. But I think we have to remind ourselves again of why we developed those technologies in the first place, and of the tremendous promises if we get it right. Of course on the one hand I think solving problems that are created by technologies, for example, existential risks — or at least some of those, they require a few kind of non-technological aspects, especially human reasoning, sense-making, and coordination.</p><p>And I’m not saying that we have to focus on one conception of the good. There are many conceptions of the good. There’s transhumanist futures, there’s cosmist futures, there’s extropian futures, and many, many more, and I think that’s fine. I don’t think we have to agree on a common conception just yet — in fact we really shouldn’t. But the point is not that we ought to settle soon, but that we have to allow into our lives again the possibility that things can be good, that good things are possible — not guaranteed, but they’re possible. I think to use technologies for good we really need a change of mindset, from pessimism to at least conditional optimism. And we need a plethora of those, right? It’s not going to be one of them.</p><p>I do think that in order to use technologies for good purposes, we really have to remind ourselves that they can be used for good, and that there are good outcomes in the first place. I genuinely think that often in our research, we put the cart before the horse in focusing solely on how catastrophic human extinction would be. I think this often misses the point that extinction is really only so bad because the potential value that could be lost is so big.</p><p><strong>Josh:</strong> If we can just make it to this point — Nick Bostrom, whose ideas a lot of The End of the World is based on, calls it technological maturity. It’s kind of a play on something that Carl Sagan said about the point we’re at now: “technological adolescence” is what Sagan called it, which is this point where we’re starting to develop this really intense, amazingly powerful technology that will one day be able to guarantee a wonderful, amazing existence for humanity, if we can survive to the point where we’ve mastered it safely. That’s what the next hundred or 200 or maybe 300 years stretches out ahead of us. That’s the challenge that we have in front of us. If we can make it to technological maturity, if we figure out how to make an artificial generalized intelligence that is friendly to humans, that basically exists to make sure that humanity is well cared for and taken care of, there’s just no telling what we’ll be able to come up with and just how vastly improved the life of the average human would be in that situation.</p><p>We’re talking — honestly, this isn’t like some crazy far out far future idea. This is conceivably something that we could get done as humans in the next century or two or three. Even if you talk out to 1000 years, that sounds far away. But really, that’s not a very long time when you consider just how far of a lifespan humanity could have stretching out ahead of it. The stakes: that makes me, almost gives me a panic attack when I think of just how close that kind of a future is for humankind and just how close to the edge we’re walking right now in developing that very same technology.</p><p><strong>Max:</strong> The way I see the future of technology as we go towards artificial general intelligence, and perhaps beyond — it could totally make life the master of its own destiny, which makes this a very important time to stop and think what do we want this destiny to be? The more clear and positive vision we can formulate, I think the more likely it is we’re going to get that destiny.</p><p><strong>Allison:</strong> We often seem to think that rather than optimizing for good outcomes, we should aim for maximizing the probability of an okay outcome, but I think for many people it’s more motivational to act on a positive vision, rather than one that is steered by risks only. To be for something rather than against something. To work toward a grand goal, rather than an outcome in which survival is success. I think a good strategy may be to focus on good outcomes.</p><p><strong>Ariel:</strong> I think it’s incredibly important to remember all of the things that we are hopeful for for the future, because these are the precise reasons that we’re trying to prevent the existential risks, all of the ways that the future could be wonderful. So let’s talk a little bit about existential hope.</p><p><strong>Allison:</strong> The term existential hope was coined by Owen Cotton-Barratt and Toby Ord to describe the chance of something extremely good happening, as opposed to an existential risk, which is a chance of something extremely terrible occurring. Kind of like describing a eucatastrophe instead of a catastrophe. I personally really agree with this line, because I think for me really it means that you can ask yourself this question of: do you think you can save the future? I think this question may appear at first pretty grandiose, but I think it’s sometimes useful to ask yourself that question, because I think if your answer is yes then you’ll likely spend your whole life trying, and you won’t rest, and that’s a pretty big decision. So I think it’s good to consider the alternative, because if the answer is no then you perhaps may be able to enjoy the little bit of time that you have on Earth rather than trying to spend it on making a difference. But I am not sure if you could actually enjoy every blissful minute right now if you knew that there was just a slight chance that you could make a difference. I mean, could you actually really enjoy this? I don’t think so, right?</p><p>I think perhaps we fail — and we do our best, but at the final moment something comes along that makes us go extinct anyways. But I think if we imagine the opposite scenario, in which we have not tried, and it turns out that we could have done something, an idea that we may have had or a skill we may have given was missing and it’s too late, I think that’s a much worse outcome.</p><p><strong>Ariel: </strong>Is it fair for me to guess, then, that you think for most people the answer is that yes, there is something that we can do to achieve a more existential hope type future?</p><p><strong>Allison:</strong> Yeah, I think so. I think that for most people there is at least something that we can be doing if we are not solving the wrong problems. But I do also think that this question is a serious question. If the answer for yourself is no, then I think you can really try to focus on having a life that is as good as it could be right now. But I do think that if the answer is yes, and if you opt in, then I think that there’s no space any more to focus on how terrible everything is. Because we’ve just confessed to how terrible everything is, and we’ve decided that we’re still going to do it. I think that if you opt in, really, then you can take that bottle of existential angst and worries that I think is really pestering us, and put it to the side for a moment. Because that’s an area you’ve dealt with and decided we’re still going to do it.</p><p><strong>Ariel:</strong> The sentiment that’s been consistent is this idea that the best way to achieve a good future is to actually figure out what we want that future to be like and aim for it.</p><p><strong>Max:</strong> On one hand, should be a no-brainer because that’s how we think about life as individuals. Right? I often get students walking into my office at MIT for career advice, and I always ask them about their vision for the future, and they always tell me something positive. They don’t walk in there and say, “Well, maybe I’ll get murdered. Maybe I’ll get cancer. Maybe I’ll …” because they know that that’s a really ridiculous approach to career planning. Instead, they envision the positive future, their aspiring things, so that we can constructively think about the challenges, the pitfalls to be avoided, and a good strategy for getting there.</p><p>Yet, as a species, we do exactly the opposite. We go to the movies and we watch Terminator, or Blade Runner, or yet another dystopic future vision that just fills us with fear and sometimes paranoia or hypochondria, when what we really need to do, as a species, is the same thing as we need to do as individuals: envision a hopeful, inspiring future that we want to rally around. It’s a well known historical fact, right, that the secret to get more constructive collaboration is to develop a shared positive vision. Why is Silicon Valley in California and not in Uruguay or Mongolia? Well, it’s because in the 60s, JFK articulated this really inspiring vision — going to space — which lead to massive investments in stem research and gave the US the best universities in the world and these amazing high tech companies, ultimately. Came from a positive vision.</p><p>Similarly, why is Germany now unified into one country instead of fragmented into many? Or Italy? Because of a positive vision. Why are the US states working together instead of having more civil wars against each other? Because of a positive vision of how much greater we’ll be if we work together. And if we can develop a more positive vision for the future of our planet, where we collaborate and everybody wins by getting richer and better off, we’re again much more likely to get that than if everybody just keeps spending their energy and time thinking about all the ways they can get screwed by their neighbors and all the ways in which things can go wrong — causing some self fulfilling prophecy basically, where we get a future with war and destruction instead of peace and prosperity.</p><p><strong>Anders:</strong> One of the things I’m envisioning is that you can make a world where everybody’s connected but also connected on their own terms. Right now, we don’t have a choice. My smartphone gives me a lot of things but it also reports my location and a lot of little apps are sending my personal information to companies and institutions I have no clue about and I don’t trust. I think one important technology that might actually be that you do privacy-enhancing technologies. Many of the little near-field microchips we carry around, they also are indiscriminately reporting to nearby antennas what we’re doing. But you could imagine having a little personal firewall that actually blocks signals that you don’t approve of. You could actually have firewalls and ways of controlling the information leaving your smartphone or your personal space. And I think we actually need to develop that, both for security purposes but also to feel that we actually are in charge of our private lives.</p><p>Some of that privacy is a social convention. We agree on what is private and not: This is why we have certain rules about what you are allowed to do with a cell phone in a restaurant. You’re not going to have a conversation with somebody — that’s rude. And others are not supposed to listen to your restaurant conversations that you have with people in the restaurant, even though technically of course, it’s trivial. I think we are going to develop new interesting rules and new technologies to help implement these social rules.</p><p>Another area I’m really excited about is the ability to capture energy, for example, using solar collectors. Solar collectors are getting exponentially better and are becoming competitive in a lot of domains with traditional energy sources. But the most beautiful things is they can be made small, used in a distributed manner. You don’t need that big central solar farm even though it might be very effective. You can actually have little solar panels on your house or even on gadgets, if they’re energy efficient enough. That means that you both reduce the risk of a collective failure but also that you get a lot of devices that can now function independently of the grid.</p><p>Then I think we are probably going to be able to combine this to fight a lot of emergent biological threats. Right now, we still have this problem that it takes a long time to identify a new pathogen. But I think we’re going to see more and more distributed sensors that can help us identify it quickly, global networks that make the medical professional aware that something new has shown up, and hopefully also ways of very quickly brewing up vaccines in an automated manner when something new shows up.</p><p>My vision is that within one or two decades, if something nasty shows up, the next morning, everybody could essentially have a little home vaccine machine manufacture those antibodies to make you resistant against that pathogen — whether that was a bio weapon or something nature accidentally brewed up.</p><p><strong>Ariel:</strong> I never even thought about our own personalized vaccine machines. Is that something people are working on?</p><p><strong>Anders: </strong>Not that much yet.</p><p><strong>Ariel: </strong>Oh.</p><p><strong>Anders:</strong> You need to manufacture antibodies cheaply and effectively. This is going to require some fairly advanced biotechnology or nanotechnology. But it’s very foreseeable. Basically, you want to have a specialized protein printer. This is something we’re moving in the direction of. I don’t think anybody’s right now doing it but I think it’s very clearly in the path where we’re already moving.</p><p>So right now in order to make a vaccine, you need to have this very time consuming process: For example in the case of flu vaccine, you identify the virus, you multiply the virus, you inject it into chicken eggs to get the antibodies and the antigens, you develop a vaccine, and if you did it all right, you have a vaccine out in a few months just in time for the winter flu — and hopefully it was for the version of the flu that was actually making the rounds. If you were unlucky, it was a different one.</p><p>But what if you could instead take the antigen, you sequence it — that’s just going to take you a few hours — you generate all the proteins, you run it through various software and biological screens to remove the ones that don’t fit, find the ones that are likely to be good targets for immune system, automatically generate the antibodies, automatically test them out so you find which ones might be bad for patients, and then test them out. Then you might be able to make a vaccine within weeks or days.</p><p><strong>Ariel:</strong> I really like your vision for the near term future. I’m hoping that all of that comes true. Now, to end, as you look further out into the future — which you’ve clearly done a lot of — what are you most hopeful for?</p><p><strong>Anders:</strong> I’m currently working on writing a book about what I call “Grand Futures.” Assuming humanity survives and gets its act together, however we’re supposed to do that, then what? How big could the future possibly be? It turns out that the laws of physics certainly allow us to do fantastic things. We might be able to spread literally over billions of light years. Settling space is definitely physically possible, but also surviving even as a normal biological species on earth for literally hundreds of millions of years — and that’s already not stretching it. It might be that if we go post-biological, we can survive up until proton decay in somewhere north of 10³⁰ years in the future. Of course, the amount of intelligence that could be generated, human brains are probably just the start.</p><p>We could probably develop ourselves or Artificial Intelligence to think enormously bigger, enormously much more deeply, enormously more profoundly. Again, this is stuff that I can analyze. There are questions about what the meaning of these thoughts would be, how deep the emotions of the future could be, et cetera, that I cannot possibly answer. But it looks like the future could be tremendously grand, enormously much bigger, just like our own current society would strike our stone age ancestors as astonishingly wealthy, astonishingly knowledgeable and interesting.</p><p>I’m looking at: what about the stability of civilizations? Historians have been going on a lot about the decline and fall of civilizations. Does that tell us an ultimate limit on what we can plan for? Eventually I got fed up reading historians and did some statistics and got some funny conclusions. But even if our civilization lasts long, it might become something very alien over time, so how do we handle that? How do you even make a backup of your civilization?</p><p>And then of course there are questions like “how long can we survive on earth?” And “when the biosphere starts failing in about a billion years, couldn’t we fix that?” What are the environmental ethics issues surrounding that? What about settling the solar system? how do you build and maintain your Dyson sphere? Then of course there’s the stellar settlement, the intergalactic settlement, then the ultimate limits of physics. What can we say about them and in what ways could physics be really different from what we expect and what does that do for our chances?</p><p>It all leads back to this question: so, what should we be doing tomorrow? What are the near term issues? Some of them are interesting like, okay, so if the future is super grand, we should probably expect that we need to safeguard ourselves against existential risk. But we might also have risks — not just going extinct, but causing suffering and pain. And maybe there are other categories we don’t know about. I’m looking a little bit at all the unknown super important things that we don’t know about yet. How do we search for them? If we discover something that turns out to be super important, how do we coordinate mankind to handle that?</p><p>Right now, this sounds totally utopian. Would you expect all humans to get together and agree on something philosophical? That sounds really unlikely. Then again, a few centuries ago the United Nations and the internet would also sound totally absurd. The future is big — we have a lot of centuries ahead of us, hopefully.</p><p><strong>Max:</strong> When I look really far into the future, I also look really far into space and I see this vast cosmos, which is 13.8 billion years old. And most of it is, despite what the UFO enthusiasts say, is actually looking pretty dead and wasted opportunities. And if we can help life flourish not just on earth, but ultimately throughout much of this amazing universe, making it come alive and teeming with these fascinating and inspiring developments, that makes me feel really, really inspired.</p><p>This is something I hope we can contribute to, we denizens of this planet, right now, here, in our lifetime. Because I think this is the most important time and place probably in cosmic history. After 13.8 billion years on this particular planet, we’ve actually developed enough technology, almost, to either drive ourselves extinct or to create super intelligence, which can spread out into the cosmos and do either horrible things or fantastic things. More than ever, life has become the master of its own destiny.</p><p><strong>Allison:</strong> For me this pretty specific vision would really be a voluntary world, in which different entities, whether they’re AI or humans, can cooperate freely with each other to realize their interests. I do think that we don’t know where we want to end up, and we really have — if you look back 100 years, it’s not only that you wouldn’t have wanted to live there, but also many of the things that were regarded as moral back then are not regarded as moral anymore by most of us, and we can expect the same to hold true 100 years from now. I think rather than locking in any specific types of values, we ought to leave the space of possible values open.</p><p>Maybe right now you could try to do something like coherent extrapolated volition, which is, in AI safety, coined by Eliezer Yudkowsky to describe a goal function of a superintelligence that would execute your goals if you were more the person you wish you were, if we lived closer together, if we had more time to think and collaborate — so kind of a perfect version of human morality. I think that perhaps we could do something like that for humans, because we all come from the same evolutionary background. We all share a few evolutionary cornerstones, at least, that make us value family, or make us value a few others of those values, and perhaps we could do something like coherent extrapolated volition of some basic, very boiled down values that most humans would agree to. I think that may be possible, I’m not sure.</p><p>On the other hand, in a future where we succeed, at least in my version of that, we live not only with humans but with a lot of different mind architectures that don’t share our evolutionary background. For those mind architectures it’s not enough to try to do something like coherent extrapolated volition, because given that they have very different starting conditions, they will also end up valuing very different value sets. In the absence of us knowing what’s in their interests, I think really the only thing we can reasonably do is try to create a framework in which very different mind architectures can cooperate freely with each other, and engage in mutually beneficial relationships.</p><p><strong>Ariel:</strong> Honestly, I really love that your answer of what you’re looking forward to is that it’s something for everybody. I like that.</p><p><strong>Anthony:</strong> When you think about what life used to be for most humans, we really have come a long way. I mean, slavery was just fully accepted for a long time. Complete subjugation of women and sexism was just totally accepted for a really long time. Poverty was just the norm. Zero political power was the norm. We are in a place where, although imperfect, many of these things have dramatically changed; even if they’re not fully implemented; Our ideals and our beliefs of human rights and human dignity and equality have completely changed and we’ve implemented a lot of that in our society.</p><p>So what I’m hopeful about is that we can continue that process, and that the way that culture and society work 100 years from now, we would look at from now and say, “Oh my God, they really have their shit together. They have figured out how to deal with differences between people, how to strike the right balance between collective desires and individual autonomy, between freedom and constraint, and how people can feel liberated to follow their own path while not trampling on the rights of others.” These are not in principle impossible things to do, and we fail to do them right now in large part, but I would like to see our technological development be leveraged into a cultural and social development that makes all those things happen. I think that really is what it’s about.</p><p>I’m much less excited about more fancy gizmos, more financial wealth for everybody, more power to have more stuff and accomplish more and higher and higher GDP. Those are useful things, but I think they’re things toward an end, and that end is the sort of happiness and fulfillment and enlightenment of the conscious living beings that make up our world. So, when I think of a positive future, it’s very much one filled with a culture that honestly will look back on ours now and say, “Boy, they really were screwed up, and I’m glad we’ve gotten better and we still have a ways to go.” And I hope that our technology will be something that will in various ways make that happen, as technology has made possible the cultural improvements we have now.</p><p><strong>Ariel:</strong> I think as a woman I do often look back at the way technology enabled feminism to happen. We needed technology to sort of get a lot of household chores accomplished — to a certain extent, I think that helped.</p><p><strong>Anthony:</strong> There are pieces of cultural progress that don’t require technology, as we were talking about earlier, but are just made so much easier by it. Labor-saving devices helped with feminism; Just industrialization I think helped with serfdom and slavery — we didn’t have to have a huge number of people working in abject poverty and total control in order for some to have a decent lifestyle, we could spread that around. I think something similar is probably true of animal suffering and meat. It could happen without that — I mean, I fully believe that 100 years from now, or 200 years from now, people will look back at eating meat as just like a crazy thing that people used to do. It’s just the truth I think of what’s going to happen.</p><p>But it’ll be much, much easier if we have technologies that make that economically viable and easy rather than pulling teeth and a huge cultural fight and everything, which I think will be hard and long. We should be thinking about, if we had some technological magic wand, what are the social problems that we would want to solve with it, and then let’s look for that wand once we identify those problems. If we could make some social problem much better if we only had such and such technology, that’s a great thing to know, because technologies are something we’re pretty good at inventing. If they don’t violate the laws of physics, and there’s some motivation, we can often generate those things, so let’s think about what they are, what would it take to solve this sort of political informational mess where nobody knows what’s true and everybody is polarized?</p><p>That’s a social problem. It has a social solution. But there might be technologies that would be enormously helpful in making those social solutions easier. So what are those technologies? Let’s think about them. So I don’t think there’s a kind of magic bullet for a lot of these problems. But having that extra boost that makes it easier to solve the social problem I think is something we should be looking for for sure.</p><p>And there are lots of technologies that really do help — worth keeping in mind, I guess, as we spend a lot of our time worrying about the ill effects of them, and the dangers and so on. There is a reason we keep pouring all this time and money and energy and creativity into developing new technologies.</p><p><strong>Ariel:</strong> I’d like to finish with one last question for everyone, and that is: what does existential hope mean for you?</p><p><strong>Max: </strong>For me, existential hope is hoping for and envisioning a really inspiring future, and then doing everything we can to make it so.</p><p><strong>Anthony:</strong> It means that we really give ourselves the space and opportunity to continue to progress our human endeavor — our culture, our society — to build a society that really is backstopping everyone’s freedom and actualization, compassion, enlightenment, in a kind of steady, ever-inventive process. I think we don’t often give ourselves as much credit as we should for how much cultural progress we’ve really made in tandem with our technological progress.</p><p><strong>Anders: </strong>My hope for the future is that we get this enormous open-ended future. It’s going to contain strange and frightening things, but I also believe that most of it is going to be fantastic. It’s going to be roaring onward far, far, far into the long term future of the universe, probably changing a lot of the aspects of the universe.</p><p>When I use the term “existential hope,” I contrast that with existential risk. Existential risks are things that threaten to curtail our entire future, to wipe it out, to make it too much smaller than it could be. Existential hope, to me, means that maybe the future is grander than we expect. Maybe we have chances we’ve never seen. And I think we are going to be surprised by many things in the future and some of them are going to be wonderful surprises. That is the real existential hope.</p><p><strong>Gaia:</strong> When I think about existential hope, I think it’s sort of an unusual phrase. But to me it’s really about the idea of finding meaning, and the potential that each of us has to experience meaning in our lives. And I think that the idea of existential hope, and I should say, the existential part of that, is the concept that that fundamental capability is something that will continue in the very long-term and will not go away. You know, I think it’s the opposite of nihilism, it’s the opposite of the idea that everything is just meaningless and our lives don’t matter and nothing that we do matters.</p><p>If I’m feeling — if I’m questioning that, I like to go and read something like Viktor Frankl’s book Man’s Search for Meaning, which really reconnects me to these incredible, deep truths about the human spirit. That’s a book that tells the story of his time in a concentration camp at Auschwitz. And even in those circumstances, the ability that he found within himself and that he saw within people around him to be kind, and to persevere, and to really give of himself, and others to give of themselves. And there’s just something impossible, I think, to capture in language. Language is a very poor tool, in this case, to try to encapsulate the essence of what that is. I think it’s something that exists on an experiential level.</p><p><strong>Allison:</strong> For me, existential hope is really trying to choose to make a difference, knowing that success is not guaranteed, but it’s really making a difference because we simply can’t do it any other way. Because not trying is really not an option. It’s the first time in history that we’ve created the technologies for our destruction and for our ascent. I think they’re both within our hands, and we have to decide how to use them. So I think existential hope is transcending existential angst, and transcending our current limitation, rather than trying to create meaning within them, and I think it’s the adequate mindset for the time that we’re in.</p><p><strong>Ariel:</strong> And I still love this idea that existential hope means that we strive toward everyone’s personal ideal, whatever that may be. On that note, I cannot thank my guests enough for joining the show, and I also hope that this episode has left everyone listening feeling a bit more optimistic about our future. I wish you all a happy holiday and a happy new year!</p><p><em>Originally published at </em><a href="https://futureoflife.org/2018/12/21/podcast-existential-hope-in-2019-and-beyond/"><em>futureoflife.org</em></a><em> on December 21, 2018.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=aa732d59bcaf" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Genome Editing and the Future of Biowarfare: A Conversation with Dr. Piers Millett]]></title>
            <link>https://medium.com/@FLIxrisk/genome-editing-and-the-future-of-biowarfare-a-conversation-with-dr-piers-millett-3b20abf59c29?source=rss-e33e2d2a809c------2</link>
            <guid isPermaLink="false">https://medium.com/p/3b20abf59c29</guid>
            <category><![CDATA[biotechnology]]></category>
            <category><![CDATA[crispr]]></category>
            <category><![CDATA[gene-editing]]></category>
            <category><![CDATA[warfare]]></category>
            <category><![CDATA[genetics]]></category>
            <dc:creator><![CDATA[Future of Life]]></dc:creator>
            <pubDate>Fri, 12 Oct 2018 18:39:55 GMT</pubDate>
            <atom:updated>2018-10-19T18:31:17.345Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/1*GQnWwNF-JXbox8WvNiMPoA.jpeg" /></figure><p><em>By Kirsten Gronlund</em></p><p>In both <a href="https://www.dni.gov/files/documents/SASC_Unclassified_2016_ATA_SFR_FINAL.pdf">2016</a> and <a href="https://www.dni.gov/files/documents/Newsroom/Testimonies/SSCI%20Unclassified%20SFR%20-%20Final.pdf">2017</a>, genome editing made it into the annual Worldwide Threat Assessment of the US Intelligence Community. One of biotechnology’s most promising modern developments, it had now been deemed a danger to US national security — and then, after two years, it was dropped from the list again. All of which raises the question: what, exactly, is genome editing, and what can it do?</p><p>Most simply, the phrase “genome editing” represents tools and techniques that biotechnologists use to edit the genome — that is, the DNA or RNA of plants, animals, and bacteria. Though the earliest versions of genome editing technology have existed for decades, the introduction of CRISPR in 2013 <a href="https://www.genome.gov/27569223/how-does-genome-editing-work/">“brought major improvements to the speed, cost, accuracy, and efficiency of genome editing.</a>”</p><p>CRISPR, or Clustered Regularly Interspersed Short Palindromic Repeats, is actually an ancient mechanism used by bacteria to remove viruses from their DNA. In the lab, researchers have discovered they can replicate this process by creating a synthetic RNA strand that matches a target DNA sequence in an organism’s genome. The RNA strand, known as a “guide RNA,” is attached to an enzyme that can cut DNA. After the guide RNA locates the targeted DNA sequence, the enzyme cuts the genome at this location. DNA can then be removed, and new DNA can be added. CRISPR has quickly become a powerful tool for editing genomes, with research taking place in a broad range of plants and animals, <a href="http://time.com/4882855/crispr-gene-editing-human-embryo/">including humans</a>.</p><p>A significant percentage of genome editing research focuses on eliminating genetic diseases. However, with tools like CRISPR, it also becomes possible to alter a pathogen’s DNA to make it more virulent and more contagious. Other potential uses include <a href="https://www.technologyreview.com/s/600774/top-us-intelligence-official-calls-gene-editing-a-wmd-threat/">the creation of “‘killer mosquitos,’ plagues that wipe out staple crops, or even a virus that snips at people’s DNA.”</a></p><p>But does genome editing really deserve a spot among the ranks of global threats like nuclear weapons and cyber hacking? To many members of the scientific community, its inclusion felt like an overreaction. Among them was Dr. Piers Millett, a science policy and international security expert whose work focuses on biotechnology and biowarfare.</p><p>Millett wasn’t surprised that biotechnology in general made it into these reports: what he didn’t expect was for one specific tool, genome editing, to be called out. In his words: “I would personally be much more comfortable if it had been a broader sentiment to say ‘Hey, there’s a whole bunch of emerging biotechnologies that could destabilize our traditional risk equation in this space, and we need to be careful with that.’ …But calling out specifically genome editing, I still don’t fully understand any rationale behind it.”</p><p>This doesn’t mean, however, that the misuse of genome editing is not cause for concern. Even proper use of the technology often involves the genetic engineering of biological pathogens, research that could very easily be weaponized. Says Millett, “If you’re deliberately trying to create a pathogen that is deadly, spreads easily, and that we don’t have appropriate public health measures to mitigate, then that thing you create is amongst the most dangerous things on the planet.”</p><h3>Biowarfare Before Genome Editing</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/720/1*1D_POqlziIjndnDt-V9yZg.jpeg" /><figcaption>A medieval depiction of the Black Plague.</figcaption></figure><p>Developments such as CRISPR present new possibilities for biowarfare, but biological weapons caused concern long before the advent of gene editing. The <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1200679/">first recorded use of biological pathogens in warfare dates back to 600 BC</a>, when Solon, an Athenian statesman, poisoned enemy water supplies during the siege of Krissa. Many centuries later, during the 1346 AD siege of Caffa, the Mongol army catapulted plague-infested corpses into the city, which is thought to have contributed to the 14th century Black Death pandemic that wiped out up to two thirds of Europe’s population.</p><p>Though biological weapons were internationally banned by the 1925 Geneva Convention, state biowarfare programs continued and in many cases expanded during World War II and the Cold War. In 1972, as evidence of these violations mounted, 103 nations signed a treaty known as the Biological Weapons Convention (BWC). The treaty bans the creation of biological arsenals and outlaws offensive biological research, though defensive research is permissible. Each year, signatories are required to submit certain information about their biological research programs to the United Nations, and violations reported to the UN Security Council may result in an inspection.</p><p>But inspections can be vetoed by the permanent members of the Security Council, and there are no firm guidelines for enforcement. On top of this, the line that separates permissible defensive biological research from its offensive counterpart is murky and remains a subject of controversy. And though the actual numbers remain unknown, pathologist Dr. Riedel asserts that <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1200679/">“the number of state-sponsored programs [that have engaged in offensive biological weapons research] has increased significantly during the last 30 years.”</a></p><h3>Dual Use Research</h3><p>So biological warfare remains a threat, and it’s one that genome editing technology could hypothetically escalate. Genome editing falls into a category of research and technology that’s known as “dual-use” — that is, it has the potential both for beneficial advances and harmful misuses. “As an enabling technology, it enables you to do things, so it is the intent of the user that determines whether that’s a positive thing or a negative thing,” Millett explains.</p><p>And ultimately, what’s considered positive or negative is a matter of perspective. “The same activity can look positive to one group of people, and negative to another. How do we decide which one is right and who gets to make that decision?” Genome editing could be used, for example, to eradicate disease-carrying mosquitoes, an application that many would consider positive. But as Millet points out, some cultures view such blatant manipulation of the ecosystem as harmful or “sacrilegious.”</p><p>Millett believes that the most effective way to deal with dual-use research is to get the researchers engaged in the discussion. “We have traditionally treated the scientific community as part of the problem,” he says. “I think we need to move to a point where the scientific community is the key to the solution, where we’re empowering them to be the ones who identify the risks, the ones who initiate the discussion about what forms this research should take.” A good scientist, he adds, is one “who’s not only doing good research, but doing research in a good way.”</p><h3>DIY Genome Editing</h3><p>But there is a growing worry that dangerous research might be undertaken by those who are not scientists at all. There are already a number of do-it-yourself (DIY) genome editing kits on the market today, and these relatively inexpensive kits <a href="https://www.scientificamerican.com/article/mail-order-crispr-kits-allow-absolutely-anyone-to-hack-dna/">allow anyone, anywhere to edit DNA</a> using CRISPR technology. Do these kits pose a real security threat? Millett explains that risk level can be assessed based on two distinct criteria: likelihood and potential impact. Where the “greatest” risks lie will depend on the criterion.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/1*8V8WiMWopp6Wv1TLSdQhDA.jpeg" /></figure><p>“If you take risk as a factor of likelihood of impact, the most likely attacks will come from low-powered actors, but have a minimal impact and be based on traditional approaches, existing pathogens, and well characterized risks and threats,” Millett explains. DIY genome editors, for example, may be great in number but are likely unable to produce a biological agent capable of causing widespread harm.</p><p>“If you switch it around and say where are the most high impact threats going to come from, then I strongly believe that that [type of threat] requires a level of sophistication and technical competency and resources that are not easy to acquire at this point in time,” says Millett. “If you’re looking for advanced stuff: who could misuse genome editing? States would be my bet in the foreseeable future.”</p><h3>State Bioweapons Programs</h3><p>Large-scale bioweapons programs, such as those run by states, pose a double threat: there is always the possibility of accidental release alongside the potential for malicious use. Millett believes that these threats are roughly equal, a conclusion backed by <a href="http://www.gryphonscientific.com/wp-content/uploads/2015/12/Final-Gain-of-Function-Risk-Benefit-Analysis-Report-12.14.2015.pdf">a thousand page report</a> from Gryphon Scientific, a US defense contractor.</p><p>Historically, both accidental release and malicious use of biological agents have caused damage. In 1979, there was the accidental release of aerosolized anthrax from the Sverdlovsk [now Ekaterinburg] bioweapons production facility in the Soviet Union — a clogged air filter in the facility had been removed, but had not been replaced. <a href="https://www.pbs.org/wgbh/pages/frontline/shows/plague/sverdlovsk/">Ninety-four people were affected by the incident and at least 64 died, along with a number of livestock</a>. The Soviet secret police attempted a cover-up and it was not until years later that the administration admitted the cause of the outbreak.</p><p>More recently, Millett says, a US biodefense facility “failed to kill the anthrax that it sent out for various lab trials, and ended up sending out really nasty anthrax around the world.” Though no one was infected, a <a href="https://www.defense.gov/Portals/1/features/2015/0615_lab-stats/Review-Committee-Report-Final.pdf">2015 government investigation</a> revealed that “over the course of the last decade, 86 facilities in the United States and seven other countries have received low concentrations of live [anthrax] spore samples… thought to be completely inactivated.”</p><p>These incidents pale, however, in comparison with Japan’s intentional use of biological weapons during the 1930s and 40s. There is “a published history that suggests up to 30,000 people were killed in China by the Japanese biological weapons program during the lead up to World War II. And if that data is accurate, that is orders of magnitude bigger than anything else,” Millett says.</p><p>Given the near-impossibility of controlling the spread of disease, a deliberate attack may have accidental effects far beyond what was intended. The Japanese, for example, may have meant to target only a few Chinese villages, only to unwittingly trigger an epidemic. There are reports, in fact, that thousands of Japan’s own soldiers became infected during a <a href="https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1200679/">biological attack in 1941</a>.</p><p>Despite the 1972 ban on biological weapons programs, Millett believes that many countries still have the capacity to produce biological weapons. As an example, he explains that the Soviets developed “a set of research and development tools that would answer the key questions and give you all the key capabilities to make biological weapons.”</p><p>The BWC only bans offensive research, and “underneath the umbrella of a defensive program,” Millett says, “you can do a whole load of research and development to figure out what you would want to weaponize if you were going to make a weapon.” Then, all a country needs to start producing those weapons is “the capacity to scale up production very, very quickly.” The Soviets, for example, built “a set of state-based commercial infrastructure to make things like vaccines.” On a day-to-day basis, they were making things the Soviet Union needed. “But they could be very radically rebooted and repurposed into production facilities for their biological weapons program,” Millett explains. This is known as a “breakout program.”</p><p>Says Millett, “I believe there are many, many countries that are well within the scope of a breakout program … so it’s not that they necessarily at this second have a fully prepared and worked-out biological weapons program that they can unleash on the world tomorrow, but they might well have all of the building blocks they need to do that in place, and a plan for how to turn their existing infrastructure towards a weapons program if they ever needed to. These components would be permissible under current international law.”</p><h3>Biological Weapons Convention</h3><p>This unsettling reality raises questions about the efficacy of the BWC — namely, what does it do well, and what doesn’t it do well? Millett, who worked for the BWC for well over a decade, has a nuanced view.</p><p>“The very fact that we have a ban on these things is brilliant,” he says. “We’re well ahead on biological weapons than many other types of weapons systems. We only got the ban on nuclear weapons — and it was only joined by some tiny number of countries — last year. Chemical weapons, only in 1995. The ban on biological weapons is hugely important. Having a space at the international level to talk about those issues is very important.” But, he adds, “we’re rapidly reaching the end of the space that I can be positive about.”</p><p>The ban on biological weapons was motivated, at least in part, by the sense that — unlike chemical weapons — they weren’t particularly useful. Traditionally, chemical and biological weapons were dealt with together. The 1925 Geneva Protocol banned both, and the original proposal for the Biological Weapons Convention, submitted by the UK in 1969, would have dealt with both. But the chemical weapons ban was ultimately dropped from the BWC, Millett says, “because that was during Vietnam, and so there were a number of chemical agents that were being used in Vietnam that weren’t going to be banned.” Once the scope of the ban had been narrowed, however, both the US and the USSR signed on.</p><p>Millet describes the resulting document as “aspirational.” He explains,“The Biological Weapons Convention is four pages long, whereas the [1995] Chemical Weapons Convention is 200 pages long, give or take.” And the difference “is about the teeth in the treaty.”</p><p>“The BWC is…a short document that’s basically a commitment by states not to make these weapons. The Chemical Weapons Convention is an international regime with an organization, with an inspection regime intended to enforce that. Under the BWC, if you are worried about another state, you’re meant to try to resolve those concerns amicably. But if you can’t do that, we move onto Article Six of the Convention, where you report it to the Security Council. The Security Council is meant to investigate it, but of course if you’re a permanent member of the Security Council, you can veto that, so that doesn’t happen.”</p><h3>De-escalation</h3><p>One easy way that states can avoid raising suspicion is to be more transparent. As Millett puts it, “If you’re not doing naughty things, then it’s on you to demonstrate that you’re not.” This doesn’t mean revealing everything to everybody. It means finding ways to show other states that they don’t need to worry.</p><p>As an example, Millett cites the heightened security culture that developed in the US after 9/11. Following the 2001 anthrax letter attacks, as well as a large investment in US biodefense programs, an initiative was started to prevent foreigners from working in those biodefense facilities. “I’m very glad they didn’t go down that path,” says Millett, “because the greatest risk, I think, was not that a foreign national would sneak in.” Rather, “the advantage of having foreign nationals in those programs was at the international level, when country Y stands up and accuses the US of having an illicit bioweapons program hidden in its biodefense program, there are three other countries that can stand up and say, ‘Well, wait a minute. Our scientists are in those facilities. We work very closely with that program, and we see no evidence of what you’re saying.’”</p><p>Historically, secrecy surrounding bioweapons programs has led other countries to begin their own research. Before World War I, the British began exploring the use of bioweapons. The Germans were aware of this. By the onset of the war, the British had abandoned the idea, but the Germans, not knowing this, began their own bioweapons program in an attempt to keep up. By World War II, Germany no longer had a bioweapons program. But the Allies believed they still did, and the U.S. bioweapons program was born of such fears.</p><h3>What now?</h3><p>Asked if he believes genome editing is a bioweapons “game changer”, Millett says no. “I see it as an enabling technology in the short to medium term, then maybe with longer-term implications [for biowarfare], but then we’re out into the far distance of what we can reasonably talk about and predict,” he says. “Certainly for now, I think its big impact is it makes it easier, faster, cheaper, and more reliable to do things that you could do using traditional approaches.”</p><p>But as biotechnology continues to evolve, so too will biowarfare. For example, it will eventually be possible for governments to alter specific genes in their own populations. “Imagine aerosolizing a lovely genome editor that knocks out a specifically nasty gene in your population,” says Millett. “It’s a passive thing. You breathe it in [and it] retroactively alters the population[’s DNA].</p><p>A government could use such technology to knock out a gene linked to cancer or other diseases. But, Millett says, “what would happen if you came across a couple of genes that at an individual level were not going to have an impact, but at a population level were connected with something, say, like IQ?” With the help of a genome editor, a government could make their population smarter, on average, by a few IQ points.</p><p>“There’s good economic data that says that [average IQ] is … statistically important,” Millett says. “The GDP of the country will be noticeably affected if we could just get another two or three percent IQ points. There are direct national security implications of that. If, for example, Chinese citizens got smarter on average over the next couple of generations by a couple of IQ points per generation, that has national security implications for both the UK and the US.”</p><p>For now, such an endeavor remains in the realm of science fiction. But technology is evolving at a breakneck speed, and it’s more important than ever to consider the potential implications of our advancements. That said, Millett is optimistic about the future. “I think the key is the distribution of bad actors versus good actors,” he says. As long as the bad actors remain the minority, there is more reason to be excited for the future of biotechnology than there is to be afraid of it.</p><p><em>Dr. Piers Millett holds fellowships at the Future of Humanity Institute, the University of Oxford, and the Woodrow Wilson Center for International Policy and works as a consultant for the World Health Organization. He also served at the United Nations as the Deputy Head of the Biological Weapons Convention.</em></p><p><em>Originally published at </em><a href="https://futureoflife.org/2018/10/12/genome-editing-and-the-future-of-biowarfare-a-conversation-with-dr-piers-millett/"><em>futureoflife.org</em></a><em> on October 12, 2018.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3b20abf59c29" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Podcast: Martin Rees on the Prospects for Humanity: AI, Biotech, Climate Change, Overpopulation…]]></title>
            <link>https://medium.com/@FLIxrisk/podcast-martin-rees-on-the-prospects-for-humanity-ai-biotech-climate-change-overpopulation-10f055a4ce6b?source=rss-e33e2d2a809c------2</link>
            <guid isPermaLink="false">https://medium.com/p/10f055a4ce6b</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[climate-change]]></category>
            <category><![CDATA[biotechnology]]></category>
            <category><![CDATA[overpopulation]]></category>
            <category><![CDATA[existential-risk]]></category>
            <dc:creator><![CDATA[Future of Life]]></dc:creator>
            <pubDate>Thu, 11 Oct 2018 19:57:35 GMT</pubDate>
            <atom:updated>2018-10-17T19:09:46.635Z</atom:updated>
            <content:encoded><![CDATA[<h3>Podcast: Martin Rees on the Prospects for Humanity: AI, Biotech, Climate Change, Overpopulation, Cryogenics, and More</h3><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fw.soundcloud.com%2Fplayer%2F%3Furl%3Dhttps%253A%252F%252Fapi.soundcloud.com%252Ftracks%252F512943534%26show_artwork%3Dtrue&amp;url=https%3A%2F%2Fsoundcloud.com%2Ffutureoflife%2Fon-the-future-an-interview-with-martin-rees&amp;image=http%3A%2F%2Fi1.sndcdn.com%2Fartworks-000418662309-to1fac-t500x500.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=soundcloud" width="800" height="166" frameborder="0" scrolling="no"><a href="https://medium.com/media/ed1b0f0bd9a090f4dee20ad7024c4228/href">https://medium.com/media/ed1b0f0bd9a090f4dee20ad7024c4228/href</a></iframe><p>How can humanity survive the next century of climate change, a growing population, and emerging technological threats? Where do we stand now, and what steps can we take to cooperate and address our greatest existential risks?</p><p>In this special podcast episode, Ariel speaks with Martin Rees about his new book, <a href="https://press.princeton.edu/titles/13251.html"><em>On the Future: Prospects for Humanity</em></a>, which discusses humanity’s existential risks and the role that technology plays in determining our collective future. Martin is a cosmologist and space scientist based in the University of Cambridge. He is director of The Institute of Astronomy and Master of Trinity College, and he was president of The Royal Society, which is the UK’s Academy of Science, from 2005 to 2010. In 2005 he was also appointed to the UK’s House of Lords.</p><p>Topics discussed in this episode include:</p><ul><li>Why Martin remains a technical optimist even as he focuses on existential risks</li><li>The economics and ethics of climate change</li><li>How AI and automation will make it harder for Africa and the Middle East to economically develop</li><li>How high expectations for health care and quality of life put society at risk</li><li>Why growing inequality could be our most under-appreciated global risk</li><li>Martin’s view that biotechnology poses greater risk than AI</li><li>Earth’s carrying capacity and the dangers of overpopulation</li><li>Space travel and why Martin is skeptical of Elon Musk’s plan to colonize Mars</li><li>The ethics of artificial meat, life extension, and cryogenics</li><li>How intelligent life could expand into the galaxy</li><li>Why humans might be unable to answer fundamental questions about the universe</li></ul><p>Books and resources discussed in this episode include:</p><ul><li><a href="https://press.princeton.edu/titles/13251.html">On the Future: Prospects for Humanity</a> by Martin Rees</li><li><a href="https://www.amazon.com/Collapse-Societies-Choose-Succeed-Revised/dp/0143117009">Collapse</a> by Jared Diamond</li><li>The <a href="https://www.effectivealtruism.org/">Effective Altruism</a> movement and <a href="https://80000hours.org/">80,000 Hours</a></li></ul><p>You can listen to the podcast above and read the full transcript below. Check out our previous podcast episodes on <a href="https://soundcloud.com/futureoflife">SoundCloud</a>, <a href="https://itunes.apple.com/us/podcast/the-future-of-life/id1170991978?mt=2">iTunes</a>, <a href="https://play.google.com/music/listen?u=0#/ps/Iatasczlcpeyawgv5werdrxbnh4">GooglePlay</a>, and <a href="https://www.stitcher.com/podcast/future-of-life-institute/the-future-of-life-institue">Stitcher</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/1*aSxIpKiv7uB6ch5mON1sWQ.jpeg" /></figure><p><strong>Ariel: </strong>Hello, I am Ariel Conn with The Future of Life Institute. Now, our podcasts lately have dealt with artificial intelligence in some way or another, and with a few focusing on nuclear weapons, but FLI is really an organization about existential risks, and especially x-risks that are the result of human action. These cover a much broader field than just artificial intelligence.</p><p>I’m excited to be hosting a special segment of the FLI podcast with Martin Rees, who has just come out with a book that looks at the ways technology and science could impact our future both for good and bad. Martin is a cosmologist and space scientist. His research interests include galaxy formation, active galactic nuclei, black holes, gamma ray bursts, and more speculative aspects of cosmology. He’s based in Cambridge where he has been director of The Institute of Astronomy, and Master of Trinity College. He was president of The Royal Society, which is the UK’s Academy of Science, from 2005 to 2010. In 2005 he was also appointed to the UK’s House of Lords. He holds the honorary title of Astronomer Royal. He has received many international awards for his research and belongs to numerous academies, including The National Academy of Sciences, the Russian Academy, the Japan Academy, and the Pontifical Academy.</p><p>He’s on the board of The Princeton Institute for Advanced Study, and has served on many bodies connected with international collaboration and science, especially threats stemming from humanity’s ever heavier footprint on the planet and the runaway consequences of ever more powerful technologies. He’s written seven books for the general public, and his most recent book is about these threats. It’s the reason that I’ve asked him to join us today. First, Martin thank you so much for talking with me today.</p><p><strong>Martin:</strong> Good to be in touch.</p><p><strong>Ariel: </strong>Your new book is called <em>On the Future: Prospects for Humanity</em>. In his endorsement of the book Neil deGrasse Tyson says, “From climate change, to biotech, to artificial intelligence, science sits at the center of nearly all decisions that civilization confronts to assure its own survival.”</p><p>I really liked this quote, because I felt like it sums up what your book is about. Basically science and the future are too intertwined to really look at one without the other. And whether the future turns out well, or whether it turns out to be the destruction of humanity, science and technology will likely have had some role to play. First, do you agree with that sentiment? Am I accurate in that description?</p><p><strong>Martin:</strong> No, I certainly agree, and that’s truer of this century than ever before because of greater scientific knowledge we have, and the greater power to use it for good or ill, because the technologies allow tremendously advanced technologies which could be misused by a small number of people.</p><p><strong>Ariel: </strong>You’ve written in the past about how you think we have essentially a 50/50 chance of some sort of existential risk. One of the things that I noticed about this most recent book is you talk a lot about the threats, but to me it felt still like an optimistic book. I was wondering if you could talk a little bit about, this might be jumping ahead a bit, but maybe what the overall message you’re hoping that people take away is?</p><p><strong>Martin: </strong>Well, I describe myself as a technical optimist, but political pessimist because it is clear that we couldn’t be living such good lives today with seven and a half billion people on the planet if we didn’t have the technology which has been developed in the last 100 years, and clearly there’s a tremendous prospect of better technology in the future. But on the other hand what is depressing is the very big gap between the way the world could be, and the way the world actually is. In particular, even though we have the power to give everyone a decent life, the lot of the bottom billion people in the world is pretty miserable and could be alleviated a lot simply by the money owned by the 1,000 richest people in the world.</p><p>We have a very unjust society, and the politics is not optimizing the way technology is used for human benefit. My view is that it’s the politics which is an impediment to the best use of technology, and the reason this is important is that as time goes on we’re going to have a growing population which is ever more demanding of energy and resources, putting more pressure on the planet and its environment and its climate, but we are also going to have to deal with this if we are to allow people to survive and avoid some serious tipping points being crossed.</p><p>That’s the problem of the collective effect of us on the planet, but there’s another effect, which is that these new technologies, especially bio, cyber, and AI allow small groups of even individuals to have an effect by error or by design, which could cascade very broadly, even globally. This, I think, makes our society very brittle. We’re very interdependent, and on the other hand it’s easy for there to be a breakdown. That’s what depresses me, the gap between the way things could be, and the downsides if we collectively overreach ourselves, or if individuals cause disruption.</p><p><strong>Ariel:</strong> You mentioned actually quite a few things that I’m hoping to touch on as we continue to talk. I’m almost inclined, before we get too far into some of the specific topics, to bring up an issue that I personally have. It’s connected to a comment that you make in the book. I think you were talking about climate change at the time, and you say that if we heard that there was 10% chance that an asteroid would strike in 2100 people would do something about it.</p><p>We wouldn’t say, “Oh, technology will be better in the future so let’s not worry about it now.” Apparently I’m very cynical, because I think that’s exactly what we would do. And I’m curious, what makes you feel more hopeful that even with something really specific like that, we would actually do something and not just constantly postpone the problem to some future generation?</p><p><strong>Martin: </strong>Well, I agree. We might not even in that case, but the reason I gave that as a contrast to our response to climate change is that there you could imagine a really sudden catastrophe happening if the asteroid does hit, whereas the problem with climate change is really that it’s first of all, the effect is mainly going to be several decades in the future. It’s started to happen, but the really severe consequences are decades away. But also there’s an uncertainty, and it’s not a sort of sudden event we can easily visualize. It’s not at all clear therefore, how we are actually going to do something about it.</p><p>In the case of the asteroid, it would be clear what the strategy would be to try and deal with it, whereas in the case of climate there are lots of ways, and the problem is that the consequences are decades away, and they’re global. Most of the political focus obviously is on short-term worry, short-term problems, and on national or more local problems. Anything we do about climate change will have an effect which is mainly for the benefit of people in quite different parts of the world 50 years from now, and it’s hard to keep those issues up the agenda when there are so many urgent things to worry about.</p><p>I think you’re maybe right that even if there was a threat of an asteroid, there may be the same sort of torpor, and we’d fail to deal with it, but I thought that’s an example of something where it would be easier to appreciate that it would really be a disaster. In the case of the climate it’s not so obviously going to be a catastrophe that people are motivated now to start thinking about it.</p><p><strong>Ariel: </strong>I’ve heard it go both ways that either climate change is yes, obviously going to be bad but it’s not an existential risk so therefore those of us who are worried about existential risk don’t need to worry about it, but then I’ve also heard people say, “No, this could absolutely be an existential risk if we don’t prevent runaway climate change.” I was wondering if you could talk a bit about what worries you most regarding climate.</p><p><strong>Martin:</strong> First of all, I don’t think it is an existential risk, but it’s something we should worry about. One point I make in my book is that I think the debate, which makes it hard to have an agreed policy on climate change, stems not so much from differences about the science — although of course there are some complete deniers — but differences about ethics and economics. There’s some people of course who completely deny the science, but most people accept that CO2 is warming the planet, and most people accept there’s quite a big uncertainty, matter of fact a true uncertainty about how much warmer you get for a given increase in CO2.</p><p>But even among those who accept the IPCC projections of climate change, and the uncertainties therein, I think there’s a big debate, and the debate is really between people who apply a standard economic discount rate where you discount the future to a rate of, say 5%, and those who think we shouldn’t do it in this context. If you apply a 5% discount rate as you would if you were deciding whether it’s worth putting up an office building or something like that, then of course you don’t give any weight to what happens after about, say 2050.</p><p>As Bjorn Lomborg, the well-known environmentalist argues, we should therefore give a lower priority to dealing with climate change than to helping the world’s poor in other more immediate ways. He is consistent given his assumptions about the discount rate. But many of us would say that in this context we should not discount the future so heavily. We should care about the life chances of a baby born today as much as we should care about the life chances of those of us who are now middle aged and won’t be alive at the end of the century. We should also be prepared to pay an insurance premium now in order to remove or reduce the risk of the worst case climate scenarios.</p><p>I think the debates about what to do about climate change is essentially ethics. Do we want to discriminate on grounds of date of birth and not care about the life chances of those who are now babies, or are we prepared to make some sacrifices now in order to reduce a risk which they might encounter in later life?</p><p><strong>Ariel: </strong>Do you think the risks are only going to be showing up that much later? We are already seeing these really heavy storms striking. We’ve got Florence in North Carolina right now. There’s a super typhoon hit southern China and the Philippines. We had Maria, and I’m losing track of all the hurricanes that we’ve had. We’ve had these huge hurricanes over the last couple of years. We saw California and much of the west coast of the US just on flames this year. Do you think we really need to wait that long?</p><p><strong>Martin: </strong>I think it’s generally agreed that extreme weather is now happening more often as a consequence of climate change and the warming of the ocean, and that this will become a more serious trend, but by the end of the century of course it could be very serious indeed. And the main threat is of course to people in the disadvantaged parts of the world. If you take these recent events, it’s been far worse in the Philippines than in the United States because they’re not prepared for it. Their houses are more fragile, etc.</p><p><strong>Ariel:</strong> I don’t suppose you have any thoughts on how we get people to care more about others? Because it does seem to be in general that sort of worrying about myself versus worrying about other people. The richer countries are the ones who are causing more of the climate change, and it’s the poorer countries who seem to be suffering more. Then of course there’s the issue of the people who are alive now versus the people in the future.</p><p><strong>Martin:</strong> That’s right, yes. Well, I think most people do care about their children and grandchildren, and so to that extent they do care about what things will be like at the end of the century, but as you say, the extra-political problem is that the cause of the CO2 emissions is mainly what’s happened in the advanced countries, and the downside is going to be more seriously felt by those in remote parts of the world. It’s easy to overlook them, and hard to persuade people that we ought to make a sacrifice which will be mainly for their benefit.</p><p>I think incidentally that’s one of the other things that we have to ensure happens, is a narrowing of the gap between the lifestyles and the economic advantages in the advanced and the less advanced parts of the world. I think that’s going to be in everyone’s interest because if there continues to be great inequality, not only will the poorer people be more subject to threats like climate change, but I think there’s going to be massive and well-justified discontent, because unlike in the earlier generations, they’re aware of what they’re missing. They all have mobile phones, they all know what it’s like, and I think there’s going to be embitterment leading to conflict if we don’t narrow this gap, and this requires I think a sacrifice on the part of the wealthy nations to subsidize developments in these poorer countries, especially in Africa.</p><p><strong>Ariel: </strong>That sort of ties into another question that I had for you, and that is, what do you think is the most underappreciated threat that maybe isn’t quite as obvious? You mentioned the fact that we have these people in poorer countries who are able to more easily see what they’re missing out on. Inequality is a problem in and of itself, but also just that people are more aware of the inequality seems like a threat that we might not be as aware of. Are there others that you think are underappreciated?</p><p><strong>Martin:</strong> Yes. Just to go back, that threat is of course very serious because by the end of the century there might be 10 times as many people in Africa as in Europe, and of course they would then have every justification in migrating towards Europe with the result of huge disruption. We do have to care about those sorts of issues. I think there are all kinds of reasons apart from straight ethics why we should ensure that the less developed countries, especially in Africa, do have a chance to close the gap.</p><p>Incidentally, one thing which is a handicap for them is that they won’t have the route to prosperity followed by the so called “Asian tigers,” which were able to have high economic growth by undercutting the labor cost in the west. Now what’s happening is that with robotics it’s possible to, as it were, re-shore lots of manufacturing industry back to wealthy countries, and so Africa and the Middle East won’t have the same opportunity the far eastern countries did to catch up by undercutting the cost of production in the west.</p><p>This is another reason why it’s going to be a big challenge. That’s something which I think we don’t worry about enough, and need to worry about, because if the inequalities persist when everyone is able to move easily and knows exactly what they’re missing, then that’s a recipe for a very dangerous and disruptive world. I would say that is an underappreciated threat.</p><p>Another thing I would count as important is that we are as a society very brittle, and very unstable because of high expectations. I’d like to give you another example. Suppose there were to be a pandemic, not necessarily a genetically engineered terrorist one, but a natural one. Then in contrast to what happened in the 14th century when the Bubonic Plague, the Black Death, occurred and killed nearly half the people in certain towns and the rest went on fatalistically. If we had some sort of plague which affected even 1% of the population of the United States, there’d be complete social breakdown, because that would overwhelm the capacity of hospitals, and people, unless they are wealthy, would feel they weren’t getting their entitlement of healthcare. And if that was a matter of life and death, that’s a recipe for social breakdown. I think given the high expectations of people in the developed world, then we are far more vulnerable to the consequences of these breakdowns, and pandemics, and the failures of electricity grids, et cetera, than in the past when people were more robust and more fatalistic.</p><p><strong>Ariel: </strong>That’s really interesting. Is it essentially because we expect to be leading these better lifestyles, just that expectation could be our downfall if something goes wrong?</p><p><strong>Martin: </strong>That’s right. And of course, if we know that there are cures available to some disease and there’s not the hospital capacity to offer it to all the people who are afflicted with the disease, then naturally that’s a matter of life and death, and that is going to promote social breakdown. This is a new threat which is of course a downside of the fact that we can at least cure some people.</p><p><strong>Ariel: </strong>There’s two directions that I want to go with this. I’m going to start with just transitioning now to biotechnology. I want to come back to issues of overpopulation and improving healthcare in a little bit, but first I want to touch on biotech threats.</p><p>One of the things that’s been a little bit interesting for me is that when I first started at FLI three years ago we were very concerned about biotechnology. CRISPR was really big. It had just sort of exploded onto the scene. Now, three years later I’m not hearing quite as much about the biotech threats, and I’m not sure if that’s because something has actually changed, or if it’s just because at FLI I’ve become more focused on AI and therefore stuff is happening but I’m not keeping up with it. I was wondering if you could talk a bit about what some of the risks you see today are with respect to biotech?</p><p><strong>Martin:</strong> Well, let me say I think we should worry far more about bio threats than about AI in my opinion. I think as far as the bio threats are concerned, then there are these new techniques. CRISPR, of course, is a very benign technique if it’s used to remove a single damaging gene that gives you a particular disease, and also it’s less objectionable than traditional GM because it doesn’t cross the species barrier in the same way, but it does allow things like a gene drive where you make a species extinct by making it sterile.</p><p>That’s good if you’re wiping out a mosquito that carries a deadly virus, but there’s a risk of some effect which distorts the ecology and has a cascading consequence. There are risks of that kind, but more important I think there is a risk of the misuse of these techniques, and not just CRISPR, but for instance the the gain of function techniques that we used in 2011 in Wisconsin and in Holland to make influenza virus both more virulent and more transmissible, things like that which can be done in a more advanced way now I’m sure.</p><p>These are clearly potentially dangerous, even if experimenters have a good motive, then the viruses might escape, and of course they are the kinds of things which could be misused. There have, of course, been lots of meetings, you have been at some, to discuss among scientists what the guidelines should be. How can we ensure responsible innovation in these technologies? These are modeled on the famous Conference in Asilomar in the 1970s when recombinant DNA was first being discussed, and the academics who worked in that area, they agreed on a sort of cautious stance, and a moratorium on some kinds of experiments.</p><p>But now they’re trying to do the same thing, and there’s a big difference. One is that these scientists are now more global. It’s not just a few people in North America and Europe. They’re global, and there is strong commercial pressures, and they’re far more widely understood. Bio-hacking is almost a student recreation. This means, in my view, that there’s a big danger, because even if we have regulations about certain things that can’t be done because they’re dangerous, enforcing those regulations globally is going to be as hopeless as it is now to enforce the drug laws, or to enforce the tax laws globally. Something which can be done will be done by someone somewhere, whatever the regulations say, and I think this is very scary. Consequences could cascade globally.</p><p><strong>Ariel:</strong> Do you think that the threat is more likely to come from something happening accidentally, or intentionally?</p><p><strong>Martin: </strong>I don’t know. I think it could be either. Certainly it could be something accidental from gene drive, or releasing some dangerous virus, but I think if we can imagine it happening intentionally, then we’ve got to ask what sort of people might do it? Governments don’t use biological weapons because you can’t predict how they will spread and who they’d actually kill, and that would be an inhibiting factor for any terrorist group that had well-defined aims.</p><p>But my worst nightmare is some person, and there are some, who think that there are too many human beings on the planet, and if they combine that view with the mindset of extreme animal rights people, etc, they might think it would be a good thing for Gaia, for Mother Earth, to get rid of a lot of human beings. They’re the kind of people who, with access to this technology, might have no compunction in releasing a dangerous pathogen. This is the kind of thing that worries me.</p><p><strong>Ariel: </strong>I find that interesting because it ties into the other question that I wanted to ask you about, and that is the idea of overpopulation. I’ve read it both ways, that overpopulation is in and of itself something of an existential risk, or a catastrophic risk, because we just don’t have enough resources on the planet. You actually made an interesting point, I thought, in your book where you point out that we’ve been thinking that there aren’t enough resources for a long time, and yet we keep getting more people and we still have plenty of resources. I thought that was sort of interesting and reassuring.</p><p>But I do think at some point that does become an issue. At then at the same time we’re seeing this huge push, understandably, for improved healthcare, and expanding life spans, and trying to save as many lives as possible, and making those lives last as long as possible. How do you resolve those two sides of the issue?</p><p><strong>Martin: </strong>It’s true, of course, as you imply, that the population has risen double in the last 50 years, and there were doomsters who in the 1960s and ’70s thought that mass starvation by now, and there hasn’t been because food production has more than kept pace. If there are famines today, as of course there are, it’s not because of overall food shortages. It’s because of wars, or mal-distribution of money to buy the food. Up until now things have gone fairly well, but clearly there are limits to the food that can be produced on the earth.</p><p>All I would say is that we can’t really say what the carrying capacity of the earth is, because it depends so much on the lifestyle of people. As I say in the book, the world couldn’t sustainably have 2 billion people if they all lived like present day Americans, using as much energy, and burning as much fossil fuels, and eating as much beef. On the other hand you could imagine lifestyles which are very sort of austere, where the earth could carry 10, or even 20 billion people. We can’t set an upper limit, but all we can say is that given that it’s fairly clear that the population is going to rise to about 9 billion by 2050, and it may go on rising still more after that, we’ve got to ensure that the way in which the average person lives is less profligate in terms of energy and resources, otherwise there will be problems.</p><p>I think we also do what we can to ensure that after 2050 the population turns around and goes down. The base scenario is when it goes on rising as it may if people choose to have large families even when they have the choice. That could happen, and of course as you say, life extension is going to have an affect on society generally, but obviously on the overall population too. I think it would be more benign if the population of 9 billion in 2050 was a peak and it started going down after that.</p><p>And it’s not hopeless, because the actual number of births per year has already started going down. The reason the population is still going up is because more babies survive, and most of the people in the developing world are still young, and if they live as long as people in advanced countries do, then of course that’s going to increase the population even for a steady birth rate. That’s why, unless there’s a real disaster, we can’t avoid the population rising to about 9 billion.</p><p>But I think policies can have an affect on what happens after that. I think we do have to try to make people realize that having large numbers of children has negative externalities, as it were in economic jargon, and it is going to be something to put extra pressure on the world, and affects our environment in a detrimental way.</p><p><strong>Ariel: </strong>As I was reading this, especially as I was reading your section about space travel, I want to ask you about your take on whether we can just start sending people to Mars or something like that to address issues of overpopulation. As I was reading your section on that, news came out that Elon Musk and SpaceX had their first passenger for a trip around the moon, which is now scheduled for 2023, and the timing was just entertaining to me, because like I said you have a section in your book about why you don’t actually agree with Elon Musk’s plan for some of this stuff.</p><p><strong>Martin:</strong> That’s right.</p><p><strong>Ariel:</strong> I was hoping you could talk a little bit about why you’re not as big a plan of space tourism, and what you think of humanity expanding into the rest of the solar system and universe?</p><p><strong>Martin:</strong> Well, let me say that I think it’s a dangerous delusion to think we can solve the earth’s problems by escaping to Mars or elsewhere. Mass emigration is not feasible. There’s nowhere in the solar system which is as comfortable to live in as the top of Everest or the South Pole. I think the idea which was promulgated by Elon Musk and Stephen Hawking of mass emigration is, I think, a dangerous delusion. The world’s problems have to be solved here, dealing with climate change is a dawdle compared to terraforming Mars. SoI don’t think that’s true.</p><p>Now, two other things about space. The first is that the practical need for sending people into space is getting less as robots get more advanced. Everyone has seen pictures of the Curiosity Probe trundling across the surface of Mars, and maybe missing things that a geologist would notice, but future robots will be able to do much of what a human will do, and to manufacture large structures in space, et cetera, so the practical need to send people to space is going down.</p><p>On the other hand, some people may want to go simply as an adventure. It’s not really tourism, because tourism implies it’s safe and routine. It’ll be an adventure like Steve Fossett or the guy who fell supersonically from an altitude balloon. It’d be crazy people like that, and maybe this Japanese tourist is in the same style, who want to have a thrill, and I think we should cheer them on.</p><p>I think it would be good to imagine that there are a few people living on Mars, but it’s never going to be as comfortable as our Earth, and we should just cheer on people like this.</p><p>And I personally think it should be left to private money. If I was an American, I would not support the NASA space program. It’s very expensive, and it could be undercut by private companies which can afford to take higher risks than NASA could inflict on publicly funded civilians. I don’t think NASA should be doing manned space flight at all. Of course, some people would say, “Well, it’s a national aspiration, a national goal to show superpower pre-eminence by a massive space project.” That was, of course, what drove the Apollo program, and the Apollo program cost about 4% of The US federal budget. Now NASA has .6% or thereabouts. I’m old enough to remember the Apollo moon landings, and of course if you would have asked me back then, I would have expected that there might have been people on Mars within 10 or 15 years at that time.</p><p>There would have been, had the program been funded, but of course there was no motive, because the Apollo program was driven by superpower rivalry. And having beaten the Russians, it wasn’t pursued with the same intensity. It could be that the Chinese will, for prestige reasons, want to have a big national space program, and leapfrog what the Americans did by going to Mars. That could happen. Otherwise I think the only manned space flight will, and indeed should, be privately funded by adventurers prepared to go on cut price and very risky missions.</p><p>But we should cheer them on. The reason we should cheer them on is that if in fact a few of them do provide some sort of settlement on Mars, then they will be important for life’s long-term future, because whereas we are, as humans, fairly well adapted to the earth, they will be in a place, Mars, or an asteroid, or somewhere, for which they are badly adapted. Therefore they would have every incentive to use all the techniques of genetic modification, and cyber technology to adapt to this hostile environment.</p><p>A new species, perhaps quite different from humans, may emerge as progeny of those pioneers within two or three centuries. I think this is quite possible. They, of course, may download themselves to be electronic. We don’t know how it’ll happen. We all know about the possibilities of advanced intelligence in electronic form. But I think this’ll happen on Mars, or in space, and of course if we think about going further and exploring beyond our solar system, then of course that’s not really a human enterprise because of human life times being limited, but it is a goal that would be feasible if you were a near immortal electronic entity. That’s a way in which our remote descendants will perhaps penetrate beyond our solar system.</p><p><strong>Ariel: </strong>As you’re looking towards these longer term futures, what are you hopeful that we’ll be able to achieve?</p><p><strong>Martin:</strong> You say we, I think we humans will mainly want to stay on the earth, but I think intelligent life, even if it’s not out there already in space, could spread through the galaxy as a consequence of what happens when a few people who go into space and are away from the regulators adapt themselves to that environment. Of course, one thing which is very important is to be aware of different time scales.</p><p>Sometimes you hear people talk about humans watching the death of the sun in five billion years. That’s nonsense, because the timescale for biological evolution by Darwinian selection is about a million years, thousands of times shorter than the lifetime of the sun, but more importantly the time scale for this new kind of intelligent design, when we can redesign humans and make new species, that time scale is a technological time scale. It could be only a century.</p><p>It would only take one, or two, or three centuries before we have entities which are very different from human beings if they are created by genetic modification, or downloading to electronic entities. They won’t be normal humans. I think this will happen, and this of course will be a very important stage in the evolution of complexity in our universe, because we will go from the kind of complexity which has emerged by Darwinian selection, to something quite new. This century is very special, which is a century where we might be triggering or jump starting a new kind of technological evolution which could spread from our solar system far beyond, on the timescale very short compared to the time scale for Darwinian evolution and the time scale for astronomical evolution.</p><p><strong>Ariel:</strong> All right. In the book you spend a lot of time also talking about current physics theories and how those could evolve. You spend a little bit of time talking about multiverses. I was hoping you could talk a little bit about why you think understanding that is important for ensuring this hopefully better future?</p><p><strong>Martin: </strong>Well, it’s only peripherally linked to it. I put that in the book because I was thinking about, what are the challenges, not just challenges of a practical kind, but intellectual challenges? One point I make is that there are some scientific challenges which we are now confronting which may be beyond human capacity to solve, because there’s no particular reason to think that the capacity of our brains is matched to understanding all aspects of reality any more than a monkey can understand quantum theory.</p><p>It’s possible that there be some fundamental aspects of nature that humans will never understand, and they will be a challenge for post-humans. I think those challenges are perhaps more likely to be in the realm of complexity, understanding the brain for instance, than in the context of cosmology, although there are challenges in cosmology which is to understand the very early universe where we may need a new theory like string theory with extra dimensions, et cetera, and we need a theory like that in order to decide whether our big bang was the only one, or whether there were other big bangs and a kind of multiverse.</p><p>It’s possible that in 50 years from now we will have such a theory, we’ll know the answers to those questions. But it could be that there is such a theory and it’s just too hard for anyone to actually understand and make predictions from. I think these issues are relevant to the intellectual constraints on humans.</p><p><strong>Ariel: </strong>Is that something that you think, or hope, that things like more advanced artificial intelligence or however we evolve in the future, that that evolution will allow “us” to understand some of these more complex ideas?</p><p><strong>Martin:</strong> Well, I think it’s certainly possible that machines could actually, in a sense, create entities based on physics which we can’t understand. This is perfectly possible, because obviously we know they can vastly out-compute us at the moment, so it could very well be, for instance, that there is a variant of string theory which is correct, and it’s just too difficult for any human mathematician to work out. But it could be that computers could work it out, so we get some answers.</p><p>But of course, you then come up against a more philosophical question about whether competence implies comprehension, whether a computer with superhuman capabilities is necessarily going to be self-aware and conscious, or whether it is going to be just a zombie. That’s a separate question which may not affect what it can actually do, but I think it does affect how we react to the possibility that the far future will be dominated by such things.</p><p>I remember when I wrote an article in a newspaper about these possibilities, the reaction was bimodal. Some people thought, “Isn’t it great there’ll be these even deeper intellects than human beings out there,” but others who thought these might just be zombies thought it was very sad if there was no entity which could actually appreciate the beauties and wonders of nature in the way we can. It does matter, in a sense, to our perception of this far future, if we think that these entities which may be electronic rather than organic, will be conscious and will have the kind of awareness that we have and which makes us wonder at the beauty of the environment in which we’ve emerged. I think that’s a very important question.</p><p><strong>Ariel:</strong> I want to pull things back to a little bit more shorter term I guess, but still considering this idea of how technology will evolve. You mentioned that you don’t think it’s a good idea to count on going to Mars as a solution to our problems on Earth because all of our problems on Earth are still going to be easier to solve here than it is to populate Mars. I think in general we have this tendency to say, “Oh, well in the future we’ll have technology that can fix whatever issue we’re dealing with now, so we don’t need to worry about it.”</p><p>I was wondering if you could sort of comment on that approach. To what extent can we say, “Well, most likely technology will have improved and can help us solve these problems,” and to what extent is that a dangerous approach to take?</p><p><strong>Martin:</strong> Well, clearly technology has allowed us to live much better, more complex lives than we could in the past, and on the whole the net benefits outweigh the downsides, but of course there are downsides, and they stem from the fact that we have some people who are disruptive, and some people who can’t be trusted. If we had a world where everyone could trust everyone else, we could get rid of about a third of the economy I would guess, but I think the main point is that we are very vulnerable.</p><p>We have huge advances, clearly, in networking via the Internet, and computers, et cetera, and we may have the Internet of Things within a decade, but of course people worry that this opens up a new kind of even more catastrophic potential for cyber terrorism. That’s just one example, and ditto for biotech which may allow the development of pathogens which kill people of particular races, or have other effects.</p><p>There are these technologies which are developing fast, and they can be used to great benefit, but they can be misused in ways that will provide new kinds of horrors that were not available in the past. It’s by no means obvious which way things will go. Will there be a continued net benefit of technology, as I think we’ve said there as been up ’til now despite nuclear weapons, et cetera, or will at some stage the downside run ahead of the benefits.</p><p>I do worry about the latter being a possibility, particularly because of this amplification factor, the fact that it only takes a few people in order to cause disruption that could cascade globally. The world is so interconnected that we can’t really have a disaster in one region without its affecting the whole world. Jared Diamond has this book called <em>Collapse</em> where he discusses five collapses of particular civilizations, whereas other parts of the world were unaffected.</p><p>I think if we really had some catastrophe, it would affect the whole world. It wouldn’t just affect parts. That’s something which is a new downside. The stakes are getting higher as technology advances, and my book is really aimed to say that these developments are very exciting, but they pose new challenges, and I think particularly they pose challenges because a few dissidents can cause more trouble, and I think it’ll make the world harder to govern. It’ll make cities and countries harder to govern, and a stronger tension between three things we want to achieve, which is security, privacy, and liberty. I think that’s going to be a challenge for all future governments.</p><p><strong>Ariel: </strong>Reading your book I very much got the impression that it was essentially a call to action to address these issues that you just mentioned. I was curious: what do you hope that people will do after reading the book, or learning more about these issues in general?</p><p><strong>Martin: </strong>Well, first of all I hope that people can be persuaded to think long term. I mentioned that religious groups, for instance, tend to think long term, and the papal encyclical in 2015 I think had a very important effect on the opinion in Latin America, Africa, and East Asia in the lead up to the Paris Climate Conference, for instance. That’s an example where someone from outside traditional politics would have an effect.</p><p>What’s very important is that politicians will only respond to an issue if it’s prominent in the press, and prominent in their inbox, and so we’ve got to ensure that people are concerned about this. Of course, I ended the book saying, “What are the special responsibilities of scientists,” because scientists clearly have a special responsibility to ensure that their work is safe, and that the public and politicians are made aware of the implications of any discovery they make.</p><p>I think that’s important, even though they should be mindful that their expertise doesn’t extend beyond their special area. That’s a reason why scientific understanding, in a general sense, is something which really has to be universal. This is important for education, because if we want to have a proper democracy where debate about these issues rises above the level of tabloid slogans, then given that the important issues that we have to discuss involve health, energy, the environment, climate, et cetera, which have scientific aspects, then everyone has to have enough feel for those aspects to participate in a debate, and also enough feel for probabilities and statistics to be not easily bamboozled by political arguments.</p><p>I think an educated population is essential for proper democracy. Obviously that’s a platitude. But the education needs to include, to a greater extent, an understanding of the scope and limits of science and technology. I make this point at the end and hope that it will lead to a greater awareness of these issues, and of course for people in universities, we have a responsibility because we can influence the younger generation. It’s certainly the case that students and people under 30 may be alive towards the end of the century are more mindful of these concerns than the middle aged and old.</p><p>It’s very important that these activities like the Effective Altruism movement, 80,000 Hours, and these other movements among students should be encouraged, because they are going to be important in spreading an awareness of long-term concerns. Public opinion can be changed. We can see the change in attitudes to drunk driving and things like that, which have happened over a few decades, and I think perhaps we can have a more environmental sensitivity so to become regarded as sort of rather naff or tacky to waste energy, and to be extravagant in consumption.</p><p>I’m hopeful that attitudes will change in a positive way, but I’m concerned simply because the politics is getting very difficult, because with social media, panic and rumor can spread at the speed of light, and small groups can have a global effect. This makes it very, very hard to ensure that we can keep things stable given that only a few people are needed to cause massive disruption. That’s something which is new, and I think is becoming more and more serious.</p><p><strong>Ariel: </strong>We’ve been talking a lot about things that we should be worrying about. Do you think there are things that we are currently worrying about that we probably can just let go of, that aren’t as big of risks?</p><p><strong>Martin: </strong>Well, I think we need to ensure responsible innovation in all new technologies. We’ve talked a lot about bio, and we are very concerned about the misuse of cyber technology. As regards AI, of course there are a whole lot of concerns to be had. I personally think that the takeover AI would be rather slower than many of the evangelists suspect, but of course we do have to ensure that humans are not victimized by some algorithm which they can’t have explained to them.</p><p>I think there is an awareness to this, and I think that what’s being done by your colleagues at MIT has been very important in raising awareness of the need for responsible innovation and ethical application of AI, and also what your group has recognized is that the order in which things happen is very important. If some computer is developed and goes rogue, that’s bad news, whereas if we have a powerful computer which is under our control, then it may help us to deal with these other problems, the problems of the misuse of biotech, et cetera.</p><p>The order in which things happen is going to be very important, but I must say I don’t completely share these concerns about machines running away and taking over, ’cause I think there’s a difference in that, for biological evolution there’s been a drive toward intelligence being favored, but so is aggression. In the case of computers, they may drive towards greater intelligence, but it’s not obvious that that is going to be combined with aggression, because they are going to be evolving by intelligent design, not the struggle of the fittest, which is the way that we evolved.</p><p><strong>Ariel: </strong>What about concerns regarding AI just in terms of being mis-programmed, and AI just being extremely competent? Poor design on our part, poor intelligent design?</p><p><strong>Martin:</strong> Well, I think in the short term obviously there are concerns about AI making decisions that affect people, and I think most of us would say that we shouldn’t be deprived of our credit rating, or put in prison on the basis of some AI algorithm which can’t be explained to us. We are entitled to have an explanation if something is done to us against our will. That is why it is worrying if too much is going to be delegated to AI.</p><p>I also think that constraint on the development of self-driving cars, and things of that kind, is going to be constrained by the fact that these become vulnerable to hacking of various kinds. I think it’ll be a long time before we will accept a driverless car on an ordinary road. Controlled environments, yes. In particular lanes on highways, yes. In an ordinary road in a traditional city, it’s not clear that we will ever accept a driverless car. I think I’m frankly less bullish than maybe some of your colleagues about the speed at which the machines will really take over and be accepted, that we can trust ourselves to them.</p><p><strong>Ariel:</strong> As I mentioned at the start, and as you mentioned at the start, you are a techno optimist, for as much as the book is about things that could go wrong it did feel to me like it was also sort of an optimistic look at the future. What are you most optimistic about? What are you most hopeful for looking at both short term and long term, however you feel like answering that?</p><p><strong>Martin:</strong> I’m hopeful that biotech will have huge benefits for health, will perhaps extend human life spans a bit, but that’s something about which we should feel a bit ambivalent. So, I think health, and also food. If you asked me, what is one of the most benign technologies, it’s to make artificial meat, for instance. It’s clear that we can more easily feed a population of 9 billion on a vegetarian diet than on a traditional diet like Americans consume today.</p><p>To take one benign technology, I would say artificial meat is one, and more intensive farming so that we can feed people without encroaching too much on the natural part of the world. I’m optimistic about that. If we think about very long term trends then life extension is something which obviously if it happens too quickly is going to be hugely disruptive, multi-generation families, et cetera.</p><p>Also, even though we will have the capability within a century to change human beings, I think we should constrain that on earth and just let that be done by the few crazy pioneers who go away into space. But if this does happen, then as I say in the introduction to my book, it will be a real game changer in a sense. I make the point that one thing that hasn’t changed over most of human history is human character. Evidence for this is that we can read the literature written by the Greeks and Romans more than 2,000 years ago and resonate with the people, and their characters, and their attitudes and emotions.</p><p>It’s not at all clear that on some scenarios, people 200 years from now will resonate in anything other than an algorithmic sense with the attitudes we have as humans today. That will be a fundamental, and very fast change in the nature of humanity. The question is, can we do something to at least constrain the rate at which that happens, or at least constrain the way in which it happens? But it is going to be almost certainly possible to completely change human mentality, and maybe even human physique over that time scale. One has only to listen to listen to people like George Church to realize that it’s not crazy to imagine this happening.</p><p><strong>Ariel:</strong> You mentioned in the book that there’s lots of people who are interested in cryogenics, but you also talked briefly about how there are some negative effects of cryogenics, and the burden that it puts on the future. I was wondering if you could talk really quickly about that?</p><p><strong>Martin: </strong>There are some people, I know some, who have a medallion around their neck which is an injunction of, if they drop dead they should be immediately frozen, and their blood drained and replaced by liquid nitrogen, and that they should then be stored — there’s a company called Alcor in Arizona that does this — and allegedly revived at some stage when technology advanced. I find it hard to take these seriously, but they say that, well the chance may be small, but if they don’t invest this way then the chance is zero that they have a resurrection.</p><p>But I actually think that even if it worked, even if the company didn’t go bust, and sincerely maintained them for centuries and they could then be revived, I still think that what they’re doing is selfish, because they’d be revived into a world that was very different. They’d be refugees from the past, and they’d therefore be imposing an obligation on the future.</p><p>We obviously feel an obligation to look after some asylum seeker or refugee, and we might feel the same if someone had been driven out of their home in the Amazonian forest for instance, and had to find a new home, but these refugees from the past, as it were, they’re imposing a burden on future generations. I’m not sure that what they’re doing is ethical. I think it’s rather selfish.</p><p><strong>Ariel: </strong>I hadn’t thought of that aspect of it. I’m a little bit skeptical of our ability to come back.</p><p><strong>Martin:</strong> I agree. I think the chances are almost zero, even if they were stored and et cetera, one would like to see this technology tried on some animal first to see if they could freeze animals at liquid nitrogen temperatures and then revive it. I think it’s pretty crazy. Then of course, the number of people doing it is fairly small, and some of the companies doing it, there’s one in Russia, which are real ripoffs I think, and won’t survive. But as I say, even if these companies did keep going for a couple of centuries, or however long is necessary, then it’s not clear to me that it’s doing good. I also quoted this nice statement about, “What happens if we clone, and create a neanderthal? Do we put him in a zoo or send him to Harvard,” said the professor from Stanford.</p><p><strong>Ariel: </strong>Those are ethical considerations that I don’t see very often. We’re so focused on what we can do that sometimes we forget. “Okay, once we’ve done this, what happens next?”</p><p>I appreciate you being here today. Those were my questions. Was there anything else that you wanted to mention that we didn’t get into?</p><p><strong>Martin:</strong> One thing we didn’t discuss, which was a serious issue, is the limits of medical treatment, because you can make extraordinary efforts to keep people alive long before they’d have died naturally, and to keep alive babies that will never live a normal life, et cetera. Well, I certainly feel that that’s gone too far at both ends of life.</p><p>One should not devote so much effort to extreme premature babies, and allow people to die more naturally. Actually, if you asked me about predictions I’d make about the next 30 or 40 years, first more vegetarianism, secondly more euthanasia.</p><p><strong>Ariel:</strong> I support both, vegetarianism, and I think euthanasia should be allowed. I think it’s a little bit barbaric that it’s not.</p><p><strong>Martin:</strong> Yes.</p><p>I think we’ve covered quite a lot, haven’t we?</p><p><strong>Ariel:</strong> I tried to.</p><p><strong>Martin:</strong> I’d just like to mention that my book does touch a lot of bases in a fairly short book. I hope it will be read not just by scientists. It’s not really a science book, although it emphasizes how scientific ideas are what’s going to determine how our civilization evolves. I’d also like to say that for those in universities, we know it’s only interim for students, but we have universities like MIT, and my University of Cambridge, we have convening power to gather people together to address these questions.</p><p>I think the value of the centers which we have in Cambridge, and you have in MIT, are that they are groups which are trying to address these very, very big issues, these threats and opportunities. The stakes are so high that if our efforts can really reduce the risk of a disaster by one part in 10,000, we’ve more than earned our keep. I’m very supportive of our Centre for Existential Risk in Cambridge, and also the Future of Life Institute which you have at MIT.</p><p>Given the huge numbers of people who are thinking about small risks like which foods are carcinogenic, and the threats of low radiation doses, et cetera, it’s not at all inappropriate that there should be some groups who are focusing on the more extreme, albeit perhaps rather improbable threats which could affect the whole future of humanity. I think it’s very important that these groups should be encouraged and fostered, and I’m privileged to be part of them.</p><p><strong>Ariel: </strong>All right. Again, the book is <em>On the Future: Prospects for Humanity</em> by Martin Rees. I do want to add, I agree with what you just said. I think this is a really nice introduction to a lot of the risks that we face. I started taking notes about the different topics that you covered, and I don’t think I got all of them, but there’s climate change, nuclear war, nuclear winter, biodiversity loss, overpopulation, synthetic biology, genome editing, bioterrorism, biological errors, artificial intelligence, cyber technology, cryogenics, and the various topics in physics, and as you mentioned the role that scientists need to play in ensuring a safe future.</p><p>I highly recommend the book as a really great introduction to the potential risks, and the hopefully much greater potential benefits that science and technology can pose for the future. Martin, thank you again for joining me today.</p><p><strong>Martin:</strong> Thank you, Ariel, for talking to me.</p><p><em>[end of recorded material]</em></p><p><em>Originally published at </em><a href="https://futureoflife.org/2018/10/11/podcast-martin-rees-on-the-prospects-for-humanity-ai-biotech-climate-change-overpopulation-cryogenics-and-more/"><em>futureoflife.org</em></a><em> on October 11, 2018.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=10f055a4ce6b" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Cognitive Biases and AI Value Alignment: An Interview with Owain Evans]]></title>
            <link>https://medium.com/@FLIxrisk/cognitive-biases-and-ai-value-alignment-an-interview-with-owain-evans-240831e9b74e?source=rss-e33e2d2a809c------2</link>
            <guid isPermaLink="false">https://medium.com/p/240831e9b74e</guid>
            <category><![CDATA[cognitive-bias]]></category>
            <category><![CDATA[research]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ai-safety]]></category>
            <dc:creator><![CDATA[Future of Life]]></dc:creator>
            <pubDate>Mon, 08 Oct 2018 14:04:30 GMT</pubDate>
            <atom:updated>2018-10-09T19:15:37.155Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1021/1*AMbG7gnAFMnUGFpj0ShW7g.jpeg" /></figure><p><em>By Tucker Davey</em></p><p>At the core of AI safety, lies the value alignment problem: how can we teach artificial intelligence systems to act in accordance with human goals and values?</p><p>Many researchers interact with AI systems to teach them human values, using techniques like inverse reinforcement learning (IRL). In theory, with IRL, an AI system can learn what humans value and how to best assist them by observing human behavior and receiving human feedback.</p><p>But human behavior doesn’t always reflect human values, and human feedback is often biased. We say we want healthy food when we’re relaxed, but then we demand greasy food when we’re stressed. Not only do we often fail to live according to our values, but many of our values contradict each other. We value getting eight hours of sleep, for example, but we regularly sleep less because we also value working hard, caring for our children, and maintaining healthy relationships.</p><p>AI systems may be able to learn a lot by observing humans, but because of our inconsistencies, some researchers worry that systems trained with IRL will be fundamentally unable to distinguish between value-aligned and misaligned behavior. This could become especially dangerous as AI systems become more powerful: inferring the wrong values or goals from observing humans could lead these systems to adopt harmful behavior.</p><h3>Distinguishing Biases and Values</h3><p>Owain Evans, a researcher at the <a href="https://www.fhi.ox.ac.uk/">Future of Humanity Institute</a>, and Andreas Stuhlmüller, president of the research non-profit <a href="https://ought.org/">Ought</a>, have explored the limitations of IRL in teaching human values to AI systems. In particular, their research exposes how cognitive biases make it difficult for AIs to learn human preferences through interactive learning.</p><p>Evans elaborates: “We want an agent to pursue some set of goals, and we want that set of goals to coincide with human goals. The question then is, if the agent just gets to watch humans and try to work out their goals from their behavior, how much are biases a problem there?”</p><p>In some cases, AIs will be able to understand patterns of common biases. Evans and Stuhlmüller discuss the psychological literature on biases in their paper, <a href="https://arxiv.org/abs/1512.05832">Learning the Preferences of Ignorant, Inconsistent Agents</a>, and in their online book, <a href="https://agentmodels.org/">agentmodels.org</a>. An example of a common pattern discussed in agentmodels.org is “time inconsistency.” Time inconsistency is the idea that people’s values and goals change depending on when you ask them. In other words, “<a href="https://agentmodels.org/chapters/5a-time-inconsistency.html">there is an inconsistency</a> between what you prefer your future self to do and what your future self prefers to do.”</p><p>Examples of time inconsistency are everywhere. For one, most people value waking up early and exercising if you ask them before bed. But come morning, when it’s cold and dark out and they didn’t get those eight hours of sleep, they often value the comfort of their sheets and the virtues of relaxation. From waking up early to avoiding alcohol, eating healthy, and saving money, humans tend to expect more from their future selves than their future selves are willing to do.</p><p>With systematic, predictable patterns like time inconsistency, IRL could make progress with AI systems. But often our biases aren’t so clear. According to Evans, deciphering which actions coincide with someone’s values and which actions spring from biases is difficult or even impossible in general.</p><p>“Suppose you promised to clean the house but you get a last minute offer to party with a friend and you can’t resist,” he suggests. “Is this a bias, or your value of living for the moment? This is a problem for using only inverse reinforcement learning to train an AI — how would it decide what are biases and values?”</p><h3>Learning the “Correct” Values</h3><p>Despite this conundrum, understanding human values and preferences is essential for AI systems, and developers have a very practical interest in training their machines to learn these preferences.</p><p>Already today, popular websites use AI to learn human preferences. With YouTube and Amazon, for instance, machine-learning algorithms observe your behavior and predict what you will want next. But while these recommendations are often useful, they have unintended consequences.</p><p>Consider the case of Zeynep Tufekci, an associate professor at the School of Information and Library Science at the University of North Carolina. After watching videos of Trump rallies to learn more about his voter appeal, Tufekci began seeing white nationalist propaganda and Holocaust denial videos on her “autoplay” queue. She soon realized that YouTube’s algorithm, optimized to keep users engaged, predictably suggests more extreme content as users watch more videos. This led her to call the website “<a href="https://www.nytimes.com/2018/03/10/opinion/sunday/youtube-politics-radical.html">The Great Radicalizer</a>.”</p><p>This value misalignment in YouTube algorithms foreshadows the dangers of interactive learning with more advanced AI systems. Instead of optimizing advanced AI systems to appeal to our short-term desires and our attraction to extremes, designers must be able to optimize them to understand our deeper values and enhance our lives.</p><p>Evans suggests that we will want AI systems that can reason through our decisions better than humans can, understand when we are making biased decisions, and “help us better pursue our long-term preferences.” However, this will entail that AIs suggest things that seem bad to humans on first blush.</p><p>One can imagine an AI system suggesting a brilliant, counterintuitive modification to a business plan, and the human just finds it ridiculous. Or maybe an AI recommends a slightly longer, stress-free driving route to a first date, but the anxious driver takes the faster route anyway, unconvinced.</p><p>To help humans understand AIs in these scenarios, Evans and Stuhlmüller have researched how AI systems could reason in ways that are comprehensible to humans and can ultimately improve upon human reasoning.</p><p>One method (invented by Paul Christiano) is called “amplification,” where humans use AIs to help them think more deeply about decisions. Evans explains: “You want a system that does exactly the same kind of thinking that we would, but it’s able to do it faster, more efficiently, maybe more reliably. But it should be a kind of thinking that if you broke it down into small steps, humans could understand and follow.”</p><p>This second concept is called “factored cognition” — the idea of breaking sophisticated tasks into small, understandable steps. According to Evans, it’s not clear how generally factored cognition can succeed. Sometimes humans can break down their reasoning into small steps, but often we rely on intuition, which is much more difficult to break down.</p><h3>Specifying the Problem</h3><p>Evans and Stuhlmüller have started a research project on amplification and factored cognition, but they haven’t solved the problem of human biases in interactive learning — rather, they’ve set out to precisely lay out these complex issues for other researchers.</p><p>“It’s more about showing this problem in a more precise way than people had done previously,” says Evans. “We ended up getting interesting results, but one of our results in a sense is realizing that this is very difficult, and understanding why it’s difficult.”</p><p><em>This article is part of a Future of Life series on the </em><a href="http://futureoflife.org/ai-safety-research/"><em>AI safety research grants</em></a><em>, which were funded by generous donations from Elon Musk and the Open Philanthropy Project.</em></p><p><em>Originally published at </em><a href="https://futureoflife.org/2018/10/08/cognitive-biases-ai-value-alignment-owain-evans/"><em>futureoflife.org</em></a><em> on October 8, 2018.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=240831e9b74e" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[AI Alignment Podcast: Moral Uncertainty and the Path to AI Alignment with William MacAskill]]></title>
            <link>https://medium.com/@FLIxrisk/ai-alignment-podcast-moral-uncertainty-and-the-path-to-ai-alignment-with-william-macaskill-62017156cbbd?source=rss-e33e2d2a809c------2</link>
            <guid isPermaLink="false">https://medium.com/p/62017156cbbd</guid>
            <category><![CDATA[effective-altruism]]></category>
            <category><![CDATA[ai-safety]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[ethics]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Future of Life]]></dc:creator>
            <pubDate>Mon, 17 Sep 2018 21:18:02 GMT</pubDate>
            <atom:updated>2018-09-18T17:44:55.087Z</atom:updated>
            <content:encoded><![CDATA[<iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fw.soundcloud.com%2Fplayer%2F%3Furl%3Dhttps%253A%252F%252Fapi.soundcloud.com%252Ftracks%252F501436392%26show_artwork%3Dtrue&amp;url=https%3A%2F%2Fsoundcloud.com%2Ffutureoflife%2Fmoral-uncertainty-and-the-path-to-ai-alignment-with-william-macaskill&amp;image=http%3A%2F%2Fi1.sndcdn.com%2Fartworks-000407124213-1k67xg-t500x500.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=soundcloud" width="800" height="166" frameborder="0" scrolling="no"><a href="https://medium.com/media/9d6d9ea63cc4b8b527dd4a6bf64159e0/href">https://medium.com/media/9d6d9ea63cc4b8b527dd4a6bf64159e0/href</a></iframe><p>How are we to make progress on AI alignment given moral uncertainty? What are the ideal ways of resolving conflicting value systems and views of morality among persons? How ought we to go about AI alignment given that we are unsure about our normative and metaethical theories? How should preferences be aggregated and persons idealized in the context of our uncertainty?</p><p>Moral Uncertainty and the Path to AI Alignment with William MacAskill is the fifth podcast in the new AI Alignment series, hosted by Lucas Perry. For those of you that are new, this series will be covering and exploring the AI alignment problem across a large variety of domains, reflecting the fundamentally interdisciplinary nature of AI alignment. Broadly, we will be having discussions with technical and non-technical researchers across areas such as machine learning, AI safety, governance, coordination, ethics, philosophy, and psychology as they pertain to the project of creating beneficial AI. If this sounds interesting to you, we hope that you will join in the conversations by following us or subscribing to our podcasts on <a href="https://www.youtube.com/channel/UC-rCCy3FQ-GItDimSR9lhzw">Youtube</a>, <a href="https://soundcloud.com/futureoflife/tracks">SoundCloud</a>, or your preferred podcast site/application.</p><p>If you’re interested in exploring the interdisciplinary nature of AI alignment, we suggest you take a look <a href="https://futureoflife.org/valuealignmentmap/">here at a preliminary landscape</a> which begins to map this space.</p><p>In this podcast, Lucas spoke with William MacAskill. Will is a professor of philosophy at the University of Oxford and is a co-founder of the <a href="https://www.centreforeffectivealtruism.org/">Center for Effective Altruism</a>, <a href="https://www.givingwhatwecan.org/">Giving What We Can</a>, and <a href="https://80000hours.org/">80,000 Hours</a>. Will helped to create the effective altruism movement and his writing is mainly focused on issues of normative and decision theoretic uncertainty, as well as general issues in ethics.</p><p>Topics discussed in this episode include:</p><ul><li>Will’s current normative and metaethical credences</li><li>The value of moral information and moral philosophy</li><li>A taxonomy of the AI alignment problem</li><li>How we ought to practice AI alignment given moral uncertainty</li><li>Moral uncertainty in preference aggregation</li><li>Moral uncertainty in deciding where we ought to be going as a society</li><li>Idealizing persons and their preferences</li><li>The most neglected portion of AI alignment</li></ul><p><em>In this interview we discuss ideas contained in the work of William MacAskill. You can learn more about Will’s work </em><a href="http://www.williammacaskill.com/"><em>here</em></a><em>, and follow him on social media </em><a href="https://twitter.com/willmacaskill?lang=en"><em>here</em></a><em>. You can find Gordon Worley’s post </em><a href="https://www.lesswrong.com/posts/uR3znuBnaevssYDZY/rfc-meta-ethical-uncertainty-in-agi-alignment"><em>here</em></a><em> and Rob Wiblin’s previous podcast with Will </em><a href="https://80000hours.org/podcast/episodes/will-macaskill-moral-philosophy/"><em>here</em></a><em>. You can hear more in the podcast above or read the transcript below.</em></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/920/1*xor8BB-Y4-1M-l-F1a0Fjw.jpeg" /></figure><p><strong>Lucas: </strong>Hey, everyone. Welcome back to the AI Alignment Podcast series at the Future of Life Institute. I’m Lucas Perry, and today we’ll be speaking with William MacAskill on moral uncertainty and its place in AI alignment. If you’ve been enjoying this series and finding it interesting or valuable, it’s a big help if you can share it on social media and follow us on your preferred listening platform.</p><p>Will is a professor of philosophy at the University of Oxford and is a co-founder of the Center for Effective Altruism, Giving What We Can, and 80,000 Hours. Will helped to create the effective altruism movement and his writing is mainly focused on issues of normative and decision theoretic uncertainty, as well as general issues and ethics. And so, without further ado, I give you William MacAskill.</p><p>Yeah, Will, thanks so much for coming on the podcast. It’s really great to have you here.</p><p><strong>Will:</strong> Thanks for having me on.</p><p><strong>Lucas:</strong> So, I guess we can start off. You can tell us a little bit about the work that you’ve been up to recently in terms of your work in the space of metaethics and moral uncertainty just over the past few years and how that’s been evolving.</p><p><strong>Will:</strong> Great. My PhD topic was on moral uncertainty, and I’m just putting the finishing touches on a book on this topic. The idea here is to appreciate the fact that we very often are just unsure about what we ought, morally speaking, to do. It’s also plausible that we ought to be unsure about what we ought morally to do. Ethics is a really hard subject, there’s tons of disagreement, it would be overconfident to think, “Oh, I’ve definitely figured out the correct moral view.” So my work focuses on not really the question of how unsure we should be, but instead what should we do given that we’re uncertain?</p><p>In particular, I look at the issue of whether we can apply the same sort of reasoning that we apply to uncertainty about matters of fact to matters of moral uncertainty. In particular, can we use what is known as “expected utility theory”, which is very widely accepted as at least approximately correct in empirical uncertainty. Can we apply that in the same way in the case of moral uncertainty?</p><p><strong>Lucas:</strong> Right. And so coming on here, you also have a book that you’ve been working on on moral uncertainty that is unpublished. Have you just been expanding this exploration in that book, diving deeper into that?</p><p><strong>Will:</strong> That’s right. There’s actually been very little that’s been written on the topic of moral uncertainty, at least in modern times, at least relative to its importance. I would think of this as a discipline that should be studied as much as consequentialism or contractualism or Kantianism is studied. But there’s really, in modern times, only one book that’s been written on the topic and that was written 18 years ago now, or published 18 years ago. What we want is this to be, firstly, just kind of definitive introduction to the topic, it’s co-authored with me as lead author, but co-authored with Toby Ord and Krista Bickfest, laying out both what we see as the most promising path forward in terms of addressing some of the challenges that face an account of decision-making under moral uncertainty, some of the implications of taking moral uncertainty seriously, and also just some of the unanswered questions.</p><p><strong>Lucas:</strong> Awesome. So I guess, just moving forward here, you have a podcast that you already did with Rob Wiblin: 80,000 Hours. So I guess we can sort of just avoid covering a lot of the basics here about your views on using expected utility calculous in moral reasoning and moral uncertainty in order to decide what one ought to do when one is not sure what one ought to do. People can go ahead and listen to that podcast, which I’ll provide a link to within the description.</p><p>It would also be good, just to sort of get a general sense of where your meta ethical partialities just generally right now tend to lie, so what sort of meta ethical positions do you tend to give the most credence to?</p><p><strong>Will:</strong> Okay, well that’s a very well put question ’cause, as with all things, I think it’s better to talk about degrees of belief rather than absolute belief. So normally if you ask a philosopher this question, we’ll say, “I’m a nihilist,” or “I’m a moral realist,” or something, so I think it’s better to split your credences. So I think I’m about 50/50 between nihilism or error theory and something that’s non-nihilistic.</p><p>Whereby nihilism or error theory, I just mean that any positive moral statement or normative statement or a evaluative statement. That includes, you ought to maximize happiness. Or, if you want a lot of money, you ought to become a banker. Or, pain is bad. That, on this view, all of those things are false. All positive, normative or evaluative claims are false. So it’s a very radical view. And we can talk more about that, if you’d like.</p><p>In terms of the rest of my credence, the view that I’m kind of most sympathetic towards in the sense of the one that occupies most of my mental attention is a relatively robust form of moral realism. It’s not clear whether it should be called kind of naturalist moral realism or non-naturalist moral realism, but the important aspect of it is just that goodness and badness are kind of these fundamental moral properties and are properties of experience.</p><p>The things that are of value are things that supervene on conscious states, in particular good states or bad states, and the way we know about them is just by direct experience with them. Just by being acquainted with a state like pain gives us a reason for thinking we ought to have less of this in the world. So that’s my kind of favored view in the sense it’s the one I’d be most likely to defend in the seminar room.</p><p>And then I give somewhat less credence in a couple of views. One is a view called “subjectivism” which is the idea that what you ought to do is determined in some sense by what you want to do. So the simplest view there would just be when I say, “I ought to do X.” That just means I want to do X in some way. Or a more sophisticated version would be ideal subjectivism where when I say I ought to do X, it means some very idealized version of myself would want myself to want to do X. Perhaps if I had limited amounts of knowledge and much clearer computational power and so on. I’m a little less sympathetic to that than many people I know. We’ll go into that.</p><p>And then a final view that I’m also less sympathetic towards is non-cognitivism, which would be the idea that our moral statements … So when I say, “Murder is wrong,” I’m not even attempting to express a proposition. What they’re doing is just expressing some emotion of mine, like, “Yuk. Murder. Ugh,” in the same way that when I said that, that wasn’t expressing any proposition, it was just expressing some sort of pro or negative attitude. And again, I don’t find that terribly plausible, again for reasons we can go into.</p><p><strong>Lucas:</strong> Right, so those first two views were cognitivist views, which makes them fall under sort of a semantic theory where you think that people are saying truth or false statements when they’re claiming moral facts. And the other theory in your moral realism are both metaphysical views, which I think is probably what we’ll mostly be interested here in terms of the AI alignment problem.</p><p>There are other issues in metaethics, for example having to do with semantics, as you just discussed. You feel as though you give some credence to non-cognitivism, but there are also justification views, so like issues in moral epistemology, how one can know about metaethics and why one ought to follow metaethics if metaethics has facts. Where do you sort of fall in in that camp?</p><p><strong>Will: </strong>Well, I think all of those views are quite well tied together, so what sort of moral epistemology you have depends very closely, I think, on what sort of meta-ethical view you have, and I actually think, often, is intimately related as well to what sort of view in normative ethics you have. So my preferred philosophical world view, as it were, the one I’d defend in a seminar room, is classical utilitarian in its normative view, so the only thing that matters is positive or negative mental states.</p><p>In terms of its moral epistemology, the way we access what is of value is just by experiencing it, so in just the same way we access conscious states. There are also some ways in which you can’t merely, you know, why is it that we should maximize the sum of good experiences rather than the product, or something? That’s a view that you’ve got to obtain by kind of reasoning rather than just purely from experience.</p><p>Part of my epistemology does appeal to whatever this spooky ability we have to reason about abstract affairs, but it’s the same sort of faculty that is used when we think about mathematics or set theory or other areas of philosophy. If, however, I had some different view, so supposing we were a subjectivist, well then moral epistemology looks very different. You’re actually just kind of reflecting on your own values, maybe looking at what you would actually do in different circumstances and so on, reflecting on your own preferences, and that’s the right way to come to the right kind of moral views.</p><p>There’s also another meta-ethical view called “constructivism” that I’m definitely not the best person to talk about with. But on that view, again it’s not really a realistic view, but on this view we just have a bunch of beliefs and intuitions and the correct moral view is just the best kind of systematization of those and beliefs or intuitions in the same way as you might think … Like linguistics, it is a science, but it’s fundamentally based just on what our linguistic intuitions are. It’s just kind of a systematization of them.</p><p>On that view, then, moral epistemology would be about reflecting on your own moral intuitions. You just got all of this data, which is the way things seem like to you, morally speaking, and then you’re just doing the systematization thing. So I feel like the question of moral epistemology can’t be answered in a vacuum. You’ve got to think about your meta-ethical view of the metaphysics of ethics at the same time.</p><p><strong>Lucas:</strong> I think I’m pretty interested in here, and also just poking a little bit more into that sort of 50% credence you give to your moral realist view, which is super interesting because it’s a view that people tend not to have, I guess, in the AI computer science rationality space, EA space. People tend to, I guess, have a lot of moral anti-realists in this space.</p><p>In my last podcast, I spoke with David Pearce, and he also seemed to sort of have a view like this, and I’m wondering if you can just sort of unpack yours a little bit, where he believed that suffering and pleasure disclose the in-built pleasure/pain access of the universe. Like you can think of minds as sort of objective features of the world, because they in fact are objective features of the world, and the phenomenology and experience of each person is objective in the same way that someone could objectively be experiencing redness, and in the same sense they could be objectively experiencing pain.</p><p>It seems to me, and I don’t fully understand the view, but the claim is that there are some sort of in-built quality or property to the hedonic qualia of suffering or pleasure that discloses its in-built value to that.</p><p><strong>Will:</strong> Yeah.</p><p><strong>Lucas: </strong>Could you unpack it a little bit more about the metaphysics of that and what that even means?</p><p><strong>Will: </strong>It sounds like David Pearce and I have quite similar views. I think relying heavily on the analogy with, or very close analogy with consciousness is going to help, where imagine you’re kind of a robot scientist, you don’t have any conscious experiences but you’re doing all this fancy science and so on, and then you kind of write out the book of the world, and i’m like, “hey, there’s this thing you missed out. It’s like conscious experience.” And you, the robot scientist, would say, “Wow, that’s just insane. You’re saying that some bits of matter have this first person subjective feel to them? Like, why on earth would we ever believe that? That’s just so out of whack with the naturalistic understanding of the world.” And it’s true. It just doesn’t make any sense from given what we know now. It’s a very strange phenomenon to exist in the world.</p><p><strong>Will: </strong>And so one of the arguments that motivates error theory is this idea of just, well, if values were to exist, they would just be so weird, what Mackie calls “queer”. It’s just so strange that just by a principle of Occam’s razor not adding strange things in to our ontology, we should assume they don’t exist.</p><p>But that argument would work in the same way against conscious experience, and the best response we’ve got is to say, no, but I know I’m conscious, and just tell by introspecting. I think we can run the same sort of argument when it comes to a property of consciousness as well, which is namely the goodness or badness of certain conscious experiences.</p><p>So now I just want you to go kind of totally a-theoretic. Imagine you’ve not thought about philosophy at all, or even science at all, and I was just to ask you, rip off one of your fingernails, or something. And then I say, “Is that experience bad?” And you would say yes.</p><p><strong>Lucas: </strong>Yeah, it’s bad.</p><p><strong>Will: </strong>And I would ask, how confident are you? The more confident that this pain is bad than that I even have hands, perhaps. That’s at least how it seems to be for me. So then it seems like, yeah, we’ve got this thing that we’re actually incredibly confident of which is the badness of pain, or at least the badness of pain for me, and so that’s what initially gives the case for then thinking, okay, well, that’s at least one objective moral fact that pain is bad, or at least pain is bad for me.</p><p><strong>Lucas: </strong>Right, so the step where I think that people will tend to get lost in this is when … I thought the part about Occam’s razor was very interesting. I think that most people are anti-realistic because they use Occam’s razor there and they think that what the hell would a value even be anyway in the third person objective sense? Like, that just seems really queer, as you put it. So I think people get lost at the step where the first person seems to simply have a property of badness to it.</p><p>I don’t know what that would mean if one has a naturalistic reductionist view of the world. There seems to be just like entropy, noise and quarks and maybe qualia as well. It’s not clear to me how we should think about properties of qualia and whether or not one can drive, obviously, “ought” statements about properties of qualia to normative statements, like “is” statements about the properties of qualia to “ought” statements?</p><p><strong>Will:</strong> One thing I want to be very clear on is just it definitely is the case that we have really no idea on this view. We are currently completely in the dark about some sort of explanation of how matter and forces and energy could result in goodness or badness, something that ought to be promoted. But that’s also true with conscious experience as well. We have no idea how on earth matter could result in kind of conscious experience. At the same time, it would be a mistake to start denying conscious experience.</p><p>And then we can ask, we say, okay, we don’t really know what’s going on but we accept that there’s conscious experience, and then I think if you were again just to completely pre theoretically start categorizing distant conscious experiences that we have, we’d say that some are red and some are blue, some are maybe more intense, some are kind of dimmer than others, you’d maybe classify them into sights and sounds and other sorts of experiences there.</p><p>I think also a very natural classification would be the ones that are good and the ones that are bad, and then I think when we cash that out further, I think it’s not nearly the case. I don’t think the best explanation is that when we say, oh, this is good or this is bad it means what we want or what we don’t want, but instead it’s like what we think we have reason to want or reason not to want. It seems to give us evidence for those sorts of things.</p><p><strong>Lucas:</strong> I guess my concern here is just that I worry that words like “good” and “bad” or “valuable” or “dis-valuable”, I feel some skepticism about whether or not they disclose some sort of intrinsic property of the qualia. I’m also not sure what the claim here is about the nature of and kinds of properties that qualia can have attached to them. I worry that goodness and badness might be some sort of evolutionary fiction which enhances us, enhances our fitness, but it doesn’t actually disclose some sort of intrinsic metaphysical quality or property of some kind of experience.</p><p><strong>Will:</strong> One thing I’ll say is, again, remember that I’ve got this 50% credence on error theory, so in general, all these questions, maybe this is just some evolutionary fiction, things just seem bad but they’re not actually, and so on. I actually think those are good arguments, and so that should give us confidence, some degree of confidence and this idea of just actually nothing matters at all.</p><p>But kind of underlying a lot of my views is this more general argument that if you’re unsure between two views, one in which just nothing matters at all, we’ve got no reasons for action, the other one we do have some reasons for action, then you can just ignore the one that says you’ve got no reasons for action ’cause you’re not going to do badly by its likes no matter what you do. If I were to go around shooting everybody, that wouldn’t be bad or wrong on nihilism. If I were to shoot lots of people, it wouldn’t be bad or wrong on nihilism.</p><p>So if there are arguments such as, I think an evolutionary argument that pushes us in the direction of kind of error theory, in a sense we can put them to the side, ’cause what we ought to do is just say, yeah, we take that really seriously. Give us a high credence in error theory, but now say, after all those arguments, what are the views, because most plausibly kind of bear their force.</p><p>So this is why with the kind of evolutionary worry, I’m just like, yes. But, supposing it’s the case that there actually are. Presumably conscious experiences themselves are useful in some evolutionary way that, again, we don’t really understand. I think, presumably, also good and bad experiences are useful in some evolutionary way that we don’t fully understand, perhaps because they have a tendency to motivate at least beings like us, and that in fact seems to be a key aspect of making a kind of goodness or badness statement. It’s at least somehow tied up to the idea of kind of motivation.</p><p>And then when I say ascribing a property to a conscious experience, I really just don’t mean whatever it is that we mean when we say that this experience is red seeming, this is experience is blue seeming, I mean, again, opens philosophical questions what we even mean by properties but in the same way this is bad seeming, this is good seeming.</p><p>Before I got into thinking about philosophy and naturalism and so on, would I have thought those things are kind of on a par, and I think I would’ve done, so it’s at least a pre theoretically justified view to think, yeah, there just is this axiological property of my experience.</p><p><strong>Lucas:</strong> This has made me much more optimistic. I think after my last podcast I was feeling quite depressed and nihilistic, and hearing you give this sort of non-naturalistic or naturalistic moral realist count is cheering me up a bit about the prospects of AI alignment and value in the world.</p><p><strong>Will:</strong> I mean, I think you shouldn’t get too optimistic. I’m also certainly wrong-</p><p><strong>Lucas:</strong> Yeah.</p><p><strong>Will:</strong> … sort of is my favorite view. But take any philosopher. What’s the chance that they’ve got the right views? Very low, probably.</p><p><strong>Lucas: </strong>Right, right. I think I also need to be careful here that human beings have this sort of psychological bias where we give a special metaphysical status and kind of meaning and motivation to things which have objective whatever to it. I guess there’s also some sort of motivation that I need to be mindful of that seeks out to make value objective or more meaningful and foundational in the universe.</p><p><strong>Will: </strong>Yeah. The thing that I think should make you feel optimistic, or at least motivated, is this argument that if nothing matters, it doesn’t matter that nothing matters. It just really ought not to affect what you do. You may as well act as if things do matter, and in fact we can have this project of trying to figure out if things matter, and that maybe could be an instrumental goal, which kind of is a purpose for life is to get to a place where we really can figure out if it has any meaning. I think that sort of argument can at least give one grounds for getting out of bed in the morning.</p><p><strong>Lucas: </strong>Right. I think there’s this philosophy paper that I saw, but I didn’t read, that was like, “nothing Matters, but it does matter”, with the one lower case M and then another capital case M, you know.</p><p><strong>Will:</strong> Oh, interesting.</p><p><strong>Lucas: </strong>Yeah.</p><p><strong>Will:</strong> It sounds a bit like 4:20 ethics.</p><p><strong>Lucas: </strong>Yeah, cool.</p><p>Moving on here into AI alignment. And before we get into this, I think that this is something that would also be interesting to hear you speak a little bit more about before we dive into AI alignment. What even is the value of moral information and moral philosophy, generally? Is this all just a bunch of BS or how can it be interesting and or useful in our lives, and in science and technology?</p><p><strong>Will:</strong> Okay, terrific. I mean, and this is something I write about in a paper I’m working on now and also in the book, as well.</p><p>So, yeah, I think the stereotype of the philosopher engaged in intellectual masturbation, not doing really much for the world at all, is quite a prevalent stereotype. I’ll not comment on whether that’s true for certain areas of philosophy. I think it’s definitely not true for certain areas within ethics. What is true is that philosophy is very hard, ethics is very hard. Most of the time when we’re trying to do this, we make very little progress.</p><p>If you look at the long-run history of thought in ethics and political philosophy, the influence is absolutely huge. Even just take Aristotle, Locke, Hobbes, Mill, and Marx. The influence of political philosophy and moral philosophy there, it shaped thousands of years of human history. Certainly not always for the better, sometimes for the worse, as well. So, ensuring that we get some of these ideas correct is just absolutely crucial.</p><p>Similarly, even in more recent times … Obviously not as influential as these other people, but also it’s been much less time so we can’t predict into the future, but if you consider Peter Singer as well, his ideas about the fact that we may have very strong obligations to benefit those who are distant strangers to us, or that we should treat animal welfare just on a par with human welfare, at least on some understanding of those ideas, that really has changed the beliefs and actions of, I think, probably tens of thousands of people, and often in really quite dramatic ways.</p><p>And then when we think about well, should we be doing more of this, is it merely that we’re influencing things randomly, or are we making things better or worse? Well, if we just look to the history of moral thought, we see that most people in most times have believed really atrocious things. Really morally abominable things. Endorsement of slavery, distinctions between races, subjugation of women, huge discrimination against non-heterosexual people, and, in part at least, it’s been ethical reflection that’s allowed us to break down some of those moral prejudices. And so we should presume that we have very similar moral prejudices now. We’ve made a little bit of progress, but do we have the one true theory of ethics now? I certainly think it’s very unlikely. And so we need to think more if we want to get to the actual ethical truth, if we don’t wanna be living out moral catastrophes in the same way as we would if we kept slaves, for example.</p><p><strong>Lucas:</strong> Right, I think we do want to do that, but I think that a bit later in the podcast we’ll get into whether or not that’s even possible, given economic, political, and militaristic forces acting upon the AI alignment problem and the issues with coordination and race to AGI.</p><p>Just to start to get into the AI alignment problem, I just wanna offer a little bit of context. It is implicit in the AI alignment problem, or value alignment problem, that AI needs to be aligned to some sort of ethic or set of ethics, this includes preferences or values or emotional dispositions, or whatever you might believe them to be. And so it seems that generally, in terms of moral philosophy, there are really two methods, or two methods in general, by which to arrive at an ethic. So, one is simply going to be through reason, and one is going to be through observing human behavior or artifacts, like books, movies, stories, or other things that we produce in order to infer and discover the observed preferences and ethics of people in the world.</p><p>The latter side of alignment methodologies are empirical and involves the agent interrogating and exploring the world in order to understand what the humans care about and value, as if values and ethics were simply a physical by-product of the world and of evolution. And the former is where ethics are arrived at through reason alone, and involve the AI or the AGI potentially going about ethics as a philosopher would, where one engages in moral reasoning about metaethics in order to determine what is correct. From the point of view of ethics, there is potentially only what the humans empirically do believe and then there is what we may or may not be able to arrive at through reason alone.</p><p>So, it seems that one or both of these methodologies can be used when aligning an AI system. And again, the distinction here is simply between sort of preference aggregation or empirical value learning approaches, or methods of instantiating machine ethics, reasoning, or decision-making in AI systems so they become agents of morality.</p><p>So, what I really wanna get into with you now is how metaethical uncertainty influences our decision over the methodology of value alignment. Over whether or not we are to prefer an empirical preference learning or aggregation type approach, or one which involved an imbuing of moral epistemology and ethical metacognition and reasoning into machine systems so it can discover what we ought to do. And how moral uncertainty, and metaethical moral uncertainty in particular, operates within both of these spaces once you’re committed to some view, or both of these views. And then we can get into issues and intertheoretic comparisons and how that arises here at many levels, the ideal way we should proceed if we could do what would be perfect, and again, what is actually likely to happen given race dynamics and political, economic, and militaristic forces.</p><p><strong>Will: </strong>Okay that sounds terrific. I mean, there’s a lot of cover there.</p><p>I think it might be worth me saying just maybe a couple of distinctions I think are relevant and kind of my overall view in this. So, in terms of distinction, I think within what broadly gets called the alignment problem, I think I’d like to distinguish between what I’d call the control problem, then kind of human values alignment problem, and then the actual alignment problem.</p><p>Where the control problem is just, can you get this AI to do what you want it to do? Where that’s maybe relatively narrowly construed, I want it to clean up my room, I don’t want it to put my cat in the bin, that’s kinda control problem. I think describing that as a technical problem is kind of broadly correct.</p><p>Second is then what gets called aligning AI with human values. For that, it might be the case that just having the AI pay attention to what humans actually do and infer their preferences that are revealed on that basis, maybe that’s a promising approach and so on. And that I think will become increasingly important as AI becomes larger and larger parts of the economy.</p><p>This is kind of already what we do when we vote for politicians who represent at least large chunks of the electorate. They hire economists who undertake kind of willingness-to-pay surveys and so on to work out what people want, on average. I do think that this is maybe more normatively loaded than people might often think, but at least you can understand that, just as the control problem is I have some relatively simple goal, which is, what do I want? I want this system to clean my room. How do I ensure that it actually does that without making mistakes that I wasn’t intending? This is kind of broader problem of, well you’ve got a whole society and you’ve got to aggregate their preferences for what kind of society wants and so on.</p><p>But I think, importantly, there’s this third thing which I called a minute ago, the actual alignment problem, so let’s run with that. Which is just working out what’s actually right and what’s actually wrong and what ought we to be doing. I do have a worry that because many people in the wider world, often when they start thinking philosophically they start endorsing some relatively simple, subjectivist or relativist views. They might think that answering this question of well, what do humans want, or what do people want, is just the same as answering what ought we to do? Whereas for kind of the reductio of that view, just go back a few hundred years where the question would have been, well, the white man’s alignment problem, where it’s like, “Well, what do we want, society?”, where that means white men.</p><p><strong>Lucas: </strong>Uh oh.</p><p><strong>Will: </strong>What do we want them to do? So similarly, unless you’ve got the kind of such a relativist view that you think that maybe that would have been correct back then, that’s why I wanna kind of distinguish this range of problems. And I know that you’re kind of most interested in that third thing, I think. Is that right?</p><p><strong>Lucas:</strong> Yeah, so I think I’m pretty interested in the second and the third thing, and I just wanna unpack a little bit of your distinction between the first and the second. So, the first was what you called the control problem, and you called the second just the plurality of human values and preferences and the issue of aligning to that in the broader context of the world.</p><p>It’s unclear to me how I get the AI to put a strawberry on the plate or to clean up my room and not kill my cat without the second thing haven been done, at least to me.</p><p>There is a sense at a very low level where your sort of working on technical AI alignment, which involves working on the MIRI approach with agential foundations and trying to work on a constraining optimization and corrigibility and docility and robustness and security and all of those sorts of things that people work on and the concrete problems in AI safety, stuff like that. But, it’s unclear to me where that sort of stuff is just limited to and includes the control problem, and where it begins requiring the system to be able to learn my preferences through interacting with me and thereby is already sort of participating in the second case where it’s sort of participating in AI alignment more generally, rather than being sort of like a low level controlled system.</p><p><strong>Will:</strong> Yeah, and I should say that on this side of things I’m definitely not an expert, not really the person to be talking to, but I think you’re right. There’s going to be some big, gray area or transition from systems. So there’s one that might be cleaning my room, or even let’s just say it’s playing some sort of game, unfortunately I forget the example … It was under the blog post, an example of the alignment problem in the wild, or something, from open AI. But, just a very simple example of the AIs playing a game, and you say, “Well, get as many points as possible.” And what you really want it to do is win a certain race, but what it ends up doing is driving this boat just round and round in circles because that’s the way of maximizing the number of points.</p><p><strong>Lucas: </strong>Reward hacking.</p><p><strong>Will: </strong>Reward hacking, exactly. That would be a kind of failure of control problem, that first in our sense. And then I believe there’s gonna be kind of gray areas, where perhaps it’s the certain sort of AI system where the whole point is it’s just implementing kind of what I want. And that might be very contextually determined, might depend on what my mood is of the day. For that, that might be a much, much harder problem and will involve kind of studying what I actually do and so on.</p><p>We could go into the question of whether you can solve the problem of cleaning a room without killing my cat. Whether that is possible to solve without solving much broader questions, maybe that’s not the most fruitful avenue of discussion.</p><p><strong>Lucas: </strong>So, let’s put aside this first case which involves the control problem, we’ll call it, and let’s focus on the second and the third, where again the second is defined as sort of the issue of the plurality of human values and preferences which can be observed, and then the third you described as us determining what we ought to do and tackling sort of the metaethics.</p><p><strong>Will:</strong> Yeah, just tackling the fundamental question of, “Where ought we to be headed as a society?” One just extra thing to add onto that is that’s just a general question for society to be answering. And if there are kind of fast, or even medium-speed, developments in AI, perhaps suddenly we’ve gotta start answering that question, or thinking about that question even harder in a more kind of clean way than we have before. But even if AI were to take a thousand years, we’d still need to answer that question, ’cause it’s just fundamentally the question of, “Where ought we to be heading as a society?”</p><p><strong>Lucas: </strong>Right, and so going back a little bit to the little taxonomy that I had developed earlier, it seems like your second case scenario would be sort of down to metaethical questions, which are behind and which influence the empirical issues with preference aggregation and there being plurality of values. And the third case would be, what would be arrived at through reason and, I guess, the reason of many different people.</p><p><strong>Will:</strong> Again, it’s gonna involve questions of metaethics as well where, again, on my theory that metaethics … It would actually just involve interacting with conscious experiences. And that’s a critical aspect of coming to understand what’s morally correct.</p><p><strong>Lucas:</strong> Okay, so let’s go into the second one first and then let’s go into the third one. And while we do that, it would be great if we could be mindful of problems in intertheoretic comparison and how they arise as we go through both. Does that sound good?</p><p><strong>Will:</strong> Yeah, that sounds great.</p><p><strong>Lucas:</strong> So, would you like to just sort of unpack, starting with the second view, the metaethics behind that, issues in how moral realism versus moral anti-realism will affect how the second scenario plays out, and other sorts of crucial considerations in metaethics that will affect the second scenario?</p><p><strong>Will:</strong> Yeah, so for the second scenario, which again, to be clear, is the aggregating of the variety of human preferences across a variety of contexts and so on, is that right?</p><p><strong>Lucas: </strong>Right, so that the agent can be fully autonomous and realized in the world that it is sort of an embodiment of human values and preferences, however construed.</p><p><strong>Will:</strong> Yeah, okay, so here I do think all the metaethics questions are gonna play a lot more role in the third question. So again, it’s funny, it’s very similar to the question of kind of what mainstream economists often think they’re doing when it comes to cost-benefit analysis. Let’s just even start in the individual case. Even there, it’s not a purely kind of descriptive enterprise, where, again, let’s not even talk about AI. You’re just looking out for me. You and I are friends and you want to do me a favor in some way, how do you make a decision about how to do me that favor, how to benefit me in some way? Well, you could just look at the things I do and then infer on the basis of that what my utility function is. So perhaps every morning I go and I rob a convenience store and then I buy some heroin and then I shoot up and-</p><p><strong>Lucas:</strong> Damn, Will!</p><p><strong>Will: </strong>That’s my day. Yes, it’s a confession. Yeah, you’re the first to hear it.</p><p><strong>Lucas:</strong> It’s crazy, in Oxford huh?</p><p><strong>Will: </strong>Yeah, Oxford University is wild.</p><p>You see that behavior on my part and you might therefore conclude, “Wow, well what Will really likes is heroin. I’m gonna do him a favor and buy him some heroin.” Now, that seems kind of commonsensically pretty ridiculous. Well, assuming I’m demonstrating all sorts of bad behavior that looks like it’s very bad for me, it looks like a compulsion and so on. So instead what we’re really doing is not merely maximizing the utility function that’s gone by my revealed preferences, we have some deeper idea of kind of what’s good for me or what’s bad for me.</p><p>Perhaps that comes down to just what I would want to want, or what I want myself to want to want to want. Perhaps you can do it in terms of what are called second-order, third-order preferences. What idealized Will would want … That is not totally clear. Well firstly, it’s really hard to know kind of what would idealized Will want. You’re gonna have to start doing at least a little bit of philosophy there. Because I tend to favor hedonism, I think that an idealized version of my friend would want the best possible experiences. That might be very different from what they think an idealized version of themselves would want because perhaps they have some objective list account of well-being and they think well, what they would also want is knowledge for the its own sake and appreciating beauty for its own sake and so on.</p><p>So, even there I think you’re gonna get into pretty tricky questions about what is good or bad for someone. And then after that you’ve got the question of preference aggregation, which is also really hard, both in theory and in practice. Where, do you just take strengths of preferences across absolutely everybody and then add them up? Well, firstly you might worry that you can’t actually make these comparisons of strengths of preferences between people. Certainly if you’re just looking at peoples revealed preferences, it’s really opaque how you would say if I prefer coffee to tea and you vice versa, who has the stronger preference? But perhaps we could look at behavioral facts to kind of try and at least anchor that, but it’s still then non-obvious that what we ought to do when we’re looking at everybody’s preferences is just maximize the sum rather than perhaps give some extra weighting to people who are more badly off, perhaps we give more priority to their interests. So this is kinda theoretical issues.</p><p>And then secondly, is kinda just practical issues of implementing that, where you actually need to ensure that people aren’t faking their preferences. And there’s a well known literature and voting theory that says that basically any aggregation system you have, any voting system, is going to be manipulable in some way. You’re gonna be able to get a better result for yourself, at least in some circumstances, by misrepresenting what you really want.</p><p>Again, these are kind of issues that our society already faces, but they’re gonna bite even harder when we’re thinking about delegating to artificial agents.</p><p><strong>Lucas: </strong>There’s two levels to this that you’re sort of elucidating. The first is that you can think of the AGI as being something which can do favors for everybody in humanity, so there are issues empirically and philosophically and in terms of understanding other agents about what sort of preferences should that AGI be maximizing for each individual, say being constrained by what is legal and what is generally converged upon as being good or right. And then there’s issues with preference aggregation which come up more given that we live in a resource-limited universe and world, where not all preferences can coexist and there has to be some sort of potential cancellation between different views.</p><p>And so, in terms of this higher level of preference aggregation … And I wanna step back here to metaethics and difficulties of intertheoretic comparison. It would seem that given your moral realist view, it would affect how the weighting would potentially be done. Because it seemed like before you were eluding to the fact that if your moral realist view would be true, then the way at which we could determine what we ought to do or what is good and true about morality would be through exploring the space of all possible experiences, right, so we can discover moral facts about experiences.</p><p><strong>Will:</strong> Mm-hmm (affirmative).</p><p>Lucas: And then in terms of preference aggregation, there would be people who would be right or wrong about what is good for them or the world.</p><p><strong>Will:</strong> Yeah, I guess this is, again why I wanna distinguish between these two types of value alignment problem, where on the second type, which is just kind of, “What does society want?” Societal preference aggregation. I wasn’t thinking of it as there being kind of right or wrong preferences.</p><p>In just the same way as there’s this question of just, “I want system to do X” but there’s a question of, “Do I want that?” or “How do you know that I want that?”, there’s a question of, “How do you know what society wants?” That’s a question in and of its own right that’s then separate from that third alignment issue I was raising, which then starts to bake in, well, if people have various moral preferences, views about how the world ought to be, yeah some are right and some are wrong. And no way should you give some aggregation over all those different views, because ideally you should give no weight to the ones that are wrong and if any are true, they get all the weight. It’s not really about kind of preference aggregation in that way.</p><p>Though, if you think about it as everyone is making certain sort of guess at the moral truth, then you could think of that like a kind of judgment aggregation problem. So, it might be like data or input for your kind of moral reasoning.</p><p><strong>Lucas:</strong> I think I was just sort of conceptually slicing this a tiny bit different from you. But that’s okay.</p><p>So, staying on this second view, it seems like there’s obviously going to be a lot of empirical issues and issues in understanding persons and idealized versions of themselves. Before we get in to intertheoretic comparison issues here, what is your view on coherent extrapolated volition, sort of, being the answer to this second part?</p><p><strong>Will:</strong> I don’t really know that much about it. From what I do know, it always seemed under-defined. As I understand it, the key idea is just, you take everyone’s idealized preferences in some sense, and then I think what you do is just take a sum of what everyone’s preference is. I’m personally quite in favor of the summation strategy. I think we can make interpersonal comparisons of strengths of preferences, and I think summing people’s preferences is the right approach.</p><p>We can use certain kinds of arguments that also have application in moral philosophy, like the idea of “If you didn’t know who you were going to be in society, how would you want to structure things? And if you’re a rational, self-interested agent, maximizing expected utility, then you’ll do the utilitarian aggregation function, so you’ll maximize the sum of preference strength.</p><p>But then, if we’re doing this idealized preference thing, all the devil’s going to be in the details of, “Well how are you doing this idealization?” Because, given my preferences for example, for what they are … I mean my preferences are absolutely … Certainly they’re incomplete, they’re almost certainly cyclical, who knows? Maybe there’s even some preferences I have that are areflexive of things, as well. Probably contradictory, as well, so there’s questions about what does it mean to idealize, and that’s going to be a very difficult question, and where a lot of the work is, I think.</p><p><strong>Lucas:</strong> So I guess, just two things here. What are sort of the timeline and actual real world working in relationship here, between the second problem that you’ve identified and the third problem that you’ve identified, and what is the role and work that preferences are doing here, for you, within the context of AI alignment, given that you’re sort of partial of a form of hedonistic consequentialism?</p><p><strong>Will:</strong> Okay, terrific, ’cause this is kind of important framing.</p><p>In terms of answering this alignment problem, the deep one of just where ought societies to be going, I think the key thing is to punt it. The key thing is to get us to a position where we can think about and reflect on this question, and really for a very long time, so I call this the long reflection. Perhaps it’s a period of a million years or something. We’ve got a lot of time on our hands. There’s really not the kind of scarce commodity, so there are various stages to get into that state.</p><p>The first is to reduce extinction risks down basically to zero, put us a position of kind of existential security. The second then is to start developing a society where we can reflect as much as possible and keep as many options open as possible.</p><p>Something that wouldn’t be keeping a lot of options open would be, say we’ve solved what I call the control problem, we’ve got these kind of lapdog AIs that are running the economy for us, and we just say, “Well, these are so smart, what we’re gonna do is just tell it, ‘Figure out what’s right and then do that.’” That would really not be keeping our options open. Even though I’m sympathetic to moral realism and so on, I think that would be quite a reckless thing to do.</p><p>Instead, what we want to have is something kind of … We’ve gotten to this position of real security. Maybe also along the way, we’ve fixed the various particularly bad problems of the present, poverty and so on, and now what we want to do is just keep our options open as much as possible and then kind of gradually work on improving our moral understanding where if that’s supplemented by AI system …</p><p>I think there’s tons of work that I’d love to see developing how this would actually work, but I think the best approach would be to get the artificially intelligent agents to be just doing moral philosophy, giving us arguments, perhaps creating new moral experiences that it thinks can be informative and so on, but letting the actual decision making or judgments about what is right and wrong be left up to us. Or at least have some kind of gradiated thing where we gradually transition the decision making more and more from human agents to artificial agents, and maybe that’s over a very long time period.</p><p>What I kind of think of as the control problem in that second level alignment problem, those are issues you face when you’re just addressing the question of, “Okay. Well, we’re now gonna have an AI run economy,” but you’re not yet needing to address the question of what’s actually right or wrong. And then my main thing there is just we should get ourselves into a position where we can take as long as we need to answer that question and have as many options open as possible.</p><p><strong>Lucas:</strong> I guess here given moral uncertainty and other issues, we would also want to factor in issues with astronomical waste into how long we should wait?</p><p><strong>Will:</strong> Yeah. That’s definitely informing my view, where it’s at least plausible that morality has an aggregative component, and if so, then the sheer vastness of the future may, because we’ve got half a billion to a billion years left on Earth, a hundred trillion years before the starts burn out, and then … I always forget these numbers, but I think like a hundred billion stars in the Milky Way, ten trillion galaxies.</p><p>With just vast resources at our disposal, the future could be astronomically good. It could also be astronomically bad. What we want to insure is that we get to the good outcome, and given the time scales involved, even what seem like an incredibly long delay, like a million years, is actually just very little time indeed.</p><p><strong>Lucas:</strong> In half a second I want to jump into whether or not this is actually likely to happen given race dynamics and that human beings are kind of crazy. The sort of timeline here is that we’re solving the technical control problem up into and on our way to sort of AGI and what might be superintelligence, and then we are also sort of idealizing everyone’s values and lives in a way such that they have more information and they can think more and have more free time and become idealized versions of themselves, given constraints within issues of values canceling each other out and things that we might end up just deeming to be impermissible.</p><p>After that is where this period of long reflection takes place, and sort of the dynamics and mechanics of that are seeming open questions. It seems that first comes computer science and global governance and coordination and strategy issues, and then comes long time of philosophy.</p><p><strong>Will: </strong>Yeah, then comes the million years of philosophy, so I guess not very surprising a philosopher would suggest this. Then the dynamics of the setup is an interesting question, and a super important one.</p><p>One thing you could do is just say, “Well, we’ve got ten billion people alive today, let’s say. We’re gonna divide the universe into ten billionths, so maybe that’s a thousand galaxies each or something.” And then you can trade after that point. I think that would get a pretty good outcome. There’s questions of whether you can enforce it or not into the future. There’s some arguments that you can. But maybe that’s not the optimal process, because especially if you think that “Wow! Maybe there’s actually some answer, something that is correct,” well, maybe a lot of people miss that.</p><p>I actually think if we did that and if there is some correct moral view, then I would hope that incredibly well informed people who have this vast amount of time, and perhaps intellectually augmented people and so on who have this vast amount of time to reflect would converge on that answer, and if they didn’t, then that would make me more suspicious of the idea that maybe there is a real face to the matter. But it’s still the early days we’d really want to think a lot about what goes into the setup of that kind of long reflection.</p><p><strong>Lucas:</strong> Given this account that you’ve just given about how this should play out in the long term or what it might look like, what is the actual probability do you think that this will happen given the way that the world actually is today and it’s just the game theoretic forces at work?</p><p><strong>Will:</strong> I think I’m going to be very hard pressed to give a probability. I don’t think I know even what my subjective credence is. But speaking qualitatively, I’d think it would be very unlikely that this is how it would play out.</p><p>Again, I’m like Brian and Dave in that I think if you look at just history, I do think moral forces have some influence. I wouldn’t say they’re the largest influence. I think probably randomness explains a huge amount of history, especially when you think about how certain events are just very determined by actions of individuals. Economic forces and technological forces, environmental changes are also huge as well. It is hard to think at least that it’s going to be likely that such a well orchestrated dynamic would occur. But I do think it’s possible and I think we can increase the chance of that happening by the careful actions that where people like FLI are doing at the moment.</p><p><strong>Lucas: </strong>That seems like the sort of ideal scenario, absolutely, but I also am worried that people don’t like to listen to moral philosophers or people in that potentially selfish government forces and things like that will end up taking over and controlling things, which is kind of sad for the cosmic endowment.</p><p><strong>Will: </strong>That’s exactly right. I think my chances … If there was some hard takeoff and sudden leap to artificial general intelligence, which I think is relatively unlikely, but again is possible, I think that’s probably the most scary ’cause it means that a huge amount of power is suddenly in the hands of a very small number of people potentially. You could end up with the very long run future of humanity being determined by the idiosyncratic preferences of just a small number of people, so it would be very dependent whether those people’s preferences are good or bad, with a kind of slow takeoff, so where there’s many decades in terms of development of AGI and it gradually getting incorporated into the economy.</p><p>I think there’s somewhat more hope there. Society will be a lot more prepared. It’s less likely that something very bad will happen. But my default presumption when we’re talking about multiple nations, billions of people doing something that’s very carefully coordinated is not going to happen. We have managed to do things that have involved international cooperation and amazing levels of operational expertise and coordination in the past. I think the eradication of smallpox is perhaps a good example of that. But it’s something that we don’t see very often, at least not now.</p><p><strong>Lucas: </strong>It looks like that we need to create a Peter Singer of AI safety of some other philosopher who has had a tremendous impact on politics and society to spread this sort of vision throughout the world such that it would more likely become realized. Is that potentially most likely?</p><p><strong>Will:</strong> Yeah. I think if a wide number of the political leaders, even if just political leaders of US, China, Russia, all were on board with global coordination on the issue of AI, or again, whatever other transformative technology might really upend things in the 21st century, and were on board with “How important it is that we get to this kind of period of long reflection where we can really figure out where we’re going,” then that alone would be very promising.</p><p>Then the question of just how promising is that I think depends a lot on maybe the robustness of … Even if you’re a moral realist, there’s a question of “How likely do you think it is that people will get the correct moral view?” It could be the case that it’s just this kind of strong attractor where even if you’ve got nothing as clean cut as the long reflection that I was describing, instead some really messy thing, perhaps various wars and it looks like feudal society or something, and anyone would say that civilization looks likely chaotic, maybe it’s the case that even given that, just given enough time and enough reasoning power, people will still converge on the same moral view.</p><p>I’m probably not as optimistic as that, but it’s at least a view that you could hold.</p><p><strong>Lucas: </strong>In terms of the different factors that are going into the AI alignment problem and the different levels you’ve identified, first, second, and third, which side do you think is lacking the most resources and attention right now? Are you most worried about the control problem, that first level? Or are you more worried about potential global coordination and governance stuff at the potential second level or moral philosophy stuff at the third?</p><p><strong>Will:</strong> Again, flagging … I’m sure I’m biased on this, but I’m currently by far the most worried on the third level. That’s for a couple of reasons. One is I just think the vast majority of the world are simple subjectivists or relativists, so the idea that we ought to be engaging in real moral thinking about how we use society, where we go with society, how we use our cosmic endowment as you put it, my strong default is that that question just never even really gets phrased.</p><p><strong>Lucas:</strong> You don’t think most people are theological moral realists?</p><p><strong>Will: </strong>Yeah. I guess it’s true that I’m just thinking about-</p><p><strong>Lucas:</strong> Our bubble?</p><p><strong>Will: </strong>My bubble, yeah. Well educated westerners. Most people in the world at least would say they’re theological moral realists. One thought is just that … I think my default is that some sort of relativistic will hold sway and people will just not really pay enough attention to think about what they ought to do. A second relevant thought is just I think the best possible universe is plausibly really, really good, like astronomically better than alternative extremely good universes.</p><p><strong>Lucas: </strong>Absolutely.</p><p><strong>Will:</strong> It’s also the case that if you’re … Even like slight small differences in moral view might lead you to optimize for extremely different things. Even just a toy example of preference utilitarianism vs hedonistic utilitarianism, what you might think of as two very similar views, I think in the actual world there’s not that much difference between them, because we just kind of know what makes people better off, at least approximately, improves their conscious experiences, it also is generally what they want, but when you’re kind of technologically unconstrained, it’s plausible to me that the optimal configuration of things will look really quite different between those two views. I guess I kind of think the default is that we get it very badly wrong and it will require really sustained work in order to ensure we get it right … If it’s the case that there is a right answer.</p><p><strong>Lucas:</strong> Is there anything with regards to issues in intertheoretic comparisons, or anything like that at any one of the three levels which we’ve discussed today that you feel we haven’t sufficiently covered or something that you would just like to talk about?</p><p><strong>Will:</strong> Yeah. I know that one of your listeners was asking whether I thought they were solvable even in principle, by some superintelligence, and I think they are. I think they are if other issues in moral philosophy are solvable. I think that’s particularly hard, but I think ethics in general is very hard.</p><p>I also think it is the case that whatever output we have at the end of this kind of long deliberation, again it’s unlikely we’ll get to credence 1 in a particular view, so we’ll have some distribution over different views, and we’ll want to take that into account. Maybe that means we do some kind of compromise action.</p><p>Maybe that means we just distribute our resources in proportion with our credence in different moral views. That’s again one of these really hard questions that we’ll want if at all possible to punt on and leave to people who can think about this in much more depth.</p><p>Then in terms of aggregating societal preferences, that’s more like the problem of interpersonal comparisons of preference strength, which is kind of formally isomorphic but is at least a definitely issue.</p><p><strong>Lucas:</strong> At the second and the third levels is where the intertheoretic problems are really going to be arising, and at that second level where the AGI is potentially working to idealize our values, I think there is again the open question about in the real world, whether or not there will be moral philosophers at the table or in politics or whoever has control over the AGI at that point in order to work on and think more deeply about intertheoretic comparisons of value at that level and timescale. Just thinking a little bit more about what we ought to do or what we should do realistically, given potential likely outcomes about whether or not this sort of thinking will or will not be at the table.</p><p><strong>Will:</strong> My default is just the crucial thing is to ensure that this thinking is more likely to be at the table. I think it is important to think about, “Well, what ought we to do then,” if we think it’s as very likely that things go badly wrong. Maybe it’s not the case that we should just be aiming to push for the optimal thing, but for some kind of second best strategy.</p><p>I think at the moment we should just be trying to push for the optimal thing. In particular, that’s in part because my views that a optimal universe is just so much better than even an extremely good one, that I just kind of think we ought to be really trying to maximize the chance that we can figure out what there is and then implement it. But it would be interesting to think about it more.</p><p><strong>Lucas:</strong> For sure. I guess just wrapping up here, did you ever have the chance to look at those two Lesswrong posts by Worley?</p><p><strong>Will: </strong>Yeah, I did.</p><p><strong>Lucas:</strong> Did you have any thoughts or comments on them? If people are interested you can find links in the description.</p><p><strong>Will:</strong> I read the posts, and I was very sympathetic in general to what he was thinking through. In particular the principle of philosophical conservatism. Hopefully I’ve shown that I’m very sympathetic to that, so trying to think “What are the minimal assumptions? Would this system be safe? Would this path make sense on a very, very wide array of different philosophical views?” I think the approach I’ve suggested, which is keeping our options open as much as possible and punting on the really hard questions, does satisfy that.</p><p>I think one of his posts was talking about “Should we assume moral realism or assume moral antirealism?” It seems like there our views differed a little bit, where I’m more worried that everyone’s going to assume some sort of subjectivism and relativism, and that there might be some moral truth out there that we’re missing and we never think to find it, because we decide that what we’re interested in is maximizing X, so we program agents to build X and then just go ahead with it, whereas actually the thing that we ought to have been optimizing for is Y. But broadly speaking, I think this question of trying to be as ecumenical as possible philosophically speaking makes a lot of sense.</p><p><strong>Lucas: </strong>Wonderful. Well, it’s really been a joy speaking, Will. Always a pleasure. Is there anything that you’d like to wrap up on, anywhere people can follow you or check you out on social media or anywhere else?</p><p><strong>Will:</strong> Yeah. You can follow me on Twitter @WillMacAskill if you want to read more on some of my work you can find me at williammacaskill.com</p><p><strong>Lucas: </strong>To be continued. Thanks again, Will. It’s really been wonderful.</p><p><strong>Will: </strong>Thanks so much, Lucas.</p><p><strong>Lucas:</strong> If you enjoyed this podcast, please subscribe, give it a like, or share it on your preferred social media platform. We’ll be back again soon with another episode in the AI Alignment series.</p><p><em>[end of recorded material]</em></p><p><em>Originally published at </em><a href="https://futureoflife.org/2018/09/17/moral-uncertainty-and-the-path-to-ai-alignment-with-william-macaskill/"><em>futureoflife.org</em></a><em> on September 17, 2018.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=62017156cbbd" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Making AI Safe in an Unpredictable World: An Interview with Thomas G. Dietterich]]></title>
            <link>https://medium.com/@FLIxrisk/making-ai-safe-in-an-unpredictable-world-an-interview-with-thomas-g-dietterich-e262a13967a6?source=rss-e33e2d2a809c------2</link>
            <guid isPermaLink="false">https://medium.com/p/e262a13967a6</guid>
            <category><![CDATA[machine-learning]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[self-driving-cars]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[ai-safety]]></category>
            <dc:creator><![CDATA[Future of Life]]></dc:creator>
            <pubDate>Mon, 17 Sep 2018 14:50:45 GMT</pubDate>
            <atom:updated>2018-10-03T17:10:51.613Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1000/1*Z6CmLZ6fXyfHEyH6zOV5Jg.jpeg" /></figure><p><em>By Jolene Creighton</em></p><p>Our AI systems work remarkably well in closed worlds. That’s because these environments contain a set number of variables, making the worlds perfectly known and perfectly predictable. In these micro environments, machines only encounter objects that are familiar to them. As a result, they always know how they should act and respond. Unfortunately, these same systems quickly become confused when they are deployed in the real world, as many objects aren’t familiar to them. This is a bit of a problem because, when an AI system becomes confused, the results can be deadly.</p><p>Consider, for example, a self-driving car that encounters a novel object. Should it speed up, or should it slow down? Or consider an autonomous weapon system that sees an anomaly. Should it attack, or should it power down? Each of these examples involve life-and-death decisions, and they reveal why, if we are to deploy advanced AI systems in real world environments, we <em>must</em> be confident that they will behave correctly when they encounter unfamiliar objects.</p><p><a href="http://web.engr.oregonstate.edu/~tgd/">Thomas G. Dietterich</a>, Emeritus Professor of Computer Science at Oregon State University, explains that solving this identification problem begins with ensuring that our AI systems aren’t too confident — that they recognize when they encounter a foreign object and don’t misidentify it as something that they are acquainted with. To achieve this, Dietterich asserts that we must move away from (or, at least, greatly modify) the discriminative training methods that currently dominate AI research.</p><p>However, to do that, we must first address the “open category problem.”</p><h3>Understanding the Open Category Problem</h3><p>When driving down the road, we can encounter a near infinite number of anomalies. Perhaps a violent storm will arise, and hail will start to fall. Perhaps our vision will become impeded by smoke or excessive fog. Although these encounters may be unexpected, the human brain is able to easily analyze new information and decide on the appropriate course of action — we will recognize a newspaper drifting across the road and, instead of abruptly slamming on the breaks, continue on our way.</p><p>Because of the way that they are programmed, our computer systems aren’t able to do the same.</p><p>“The way we use machine learning to create AI systems and software these days generally uses something called ‘discriminative training,’” Dietterich explains, “which implicitly assumes that the world consists of only, say, a thousand different kinds of objects.” This means that, if a machine encounters a novel object, it will assume that it <em>must</em> be one of the thousand things that it was trained on. As a result, such systems misclassify all foreign objects.</p><p>This is the “open category problem” that Dietterich and his team are attempting to solve. Specifically, they are trying to ensure that our machines don’t assume that they have encountered every possible object, but are, instead, able to reliably detect — and ultimately respond to — new categories of alien objects.</p><p>Dietterich notes that, from a practical standpoint, this means creating an anomaly detection algorithm that assigns an anomaly score to each object detected by the AI system. That score must be compared against a set threshold and, if the anomaly score exceeds the threshold, the system will need to raise an alarm. Dietterich states that, in response to this alarm, the AI system should take a pre-determined safety action. For example, a self-driving car that detects an anomaly might slow down and pull off to the side of the road.</p><h3>Creating a Theoretical Guarantee of Safety</h3><p>There are two challenges to making this method work. First, Dietterich asserts that we need good anomaly detection algorithms. Previously, in order to determine what algorithms work well, the team compared the performance of eight state-of-the-art anomaly detection algorithms on a large collection of benchmark problems.</p><p>The second challenge is to set the alarm threshold so that the AI system is guaranteed to detect a desired fraction of the alien objects, such as 99%. Dietterich says that formulating a reliable setting for this threshold is one of the most challenging research problems because there are, potentially, infinite kinds of alien objects. “The problem is that we can’t have labeled training data for <em>all</em> of the aliens. If we had such data, we would simply train the discriminative classifier on that labeled data,” Dietterich says.</p><p>To circumvent this labeling issue, the team assumes that the discriminative classifier has access to a representative sample of “query objects” that reflect the larger statistical population. Such a sample could, for example, be obtained by collecting data from cars driving on highways around the world. This sample will include some fraction of unknown objects, and the remaining objects belong to known object categories.</p><p>Notably, the data in the sample is not labeled. Instead, the AI system is given an estimate of the fraction of aliens in the sample. And by combining the information in the sample with the labeled training data that was employed to train the discriminative classifier, the team’s new algorithm can choose a good alarm threshold. If the estimated fraction of aliens is known to be an over-estimate of the true fraction, then the chosen threshold is guaranteed to detect the target percentage of aliens (i.e. 99%).</p><p>Ultimately, the above is the first method that can give a theoretical guarantee of safety for detecting alien objects, and a paper reporting the results was presented at ICML 2018. “We are able to guarantee, with high probability, that we can find 99% all of these new objects,” Dietterich says.</p><p>In the next stage of their research, Dietterich and his team plan to begin testing their algorithm in a more complex setting. Thus far, they’ve been looking primarily at classification, where the system looks at an image and classifies it. Next, they plan to move to controlling an agent, like a robot of self-driving car. “At each point in time, in order to decide what action to choose, our system will do a ‘look ahead search’ based on a learned model of the behavior of the agent and its environment. If the look ahead arrives at a state that is rated as ‘alien’ by our method, then this indicates that the agent is about to enter a part of the state space where it is not competent to choose correct actions,” Dietterich says. In response, as previously mentioned, the agent should execute a series of safety actions and request human assistance.</p><p>But what does this safety action actually consist of?</p><h3>Responding to Aliens</h3><p>Dietterich notes that, once something is identified as an anomaly and the alarm is sounded, the nature of this fall back system will depend on the machine in question, like whether the AI system is in a self-driving car or autonomous weapon.</p><p>To explain how these secondary systems operate, Dietterich turns to self-driving cars. “In the Google car, if the computers lose power, then there’s a backup system that automatically slows the car down and pulls it over to the side of the road.” However, Dietterich clarifies that stopping isn’t always the best course of action. One may assume that a car should come to a halt if an unidentified object crosses its path; however, if the unidentified object happens to be a blanket of snow on a particularly icy day, hitting the breaks gets more complicated. The system would need to factor in the icy roads, any cars that may be driving behind, and whether these cars can break in time to avoid a rear end collision.</p><p>But if we can’t predict every eventuality, how can we expect to program an AI system so that it behaves correctly and in a way that is safe?</p><p>Unfortunately, there’s no easy answer; however, Dietterich clarifies that there are some general best practices; “There’s no universal solution to the safety problem, but obviously there are some actions that are safer than others. Generally speaking, removing energy from the system is a good idea,” he says. Ultimately, Dietterich asserts that all the work related to programming safe AI really boils down to determining how we want our machines to behave under specific scenarios, and he argues that we need to rearticulate how we characterize this problem, and focus on accounting for all the factors, if we are to develop a sound approach.</p><p>Dietterich notes that “when we look at these problems, they tend to get lumped under a classification of ‘ethical decision making,’ but what they really are is problems that are incredibly complex. They depend tremendously on the context in which they are operating, the human beings, the other innovations, the other automated systems, and so on. The challenge is correctly describing how we want the system to behave and then ensuring that our implementations actually comply with those requirements.” And he concludes, “the big risk in the future of AI is the same as the big risk in any software system, which is that we build the wrong system, and so it does the wrong thing. Arthur C Clark in <em>2001: A Space Odyssey</em> had it exactly right. The Hal 9000 didn’t ‘go rogue;’ it was just doing what it had been programmed to do.”</p><p><em>This article is part of a Future of Life series on the </em><a href="http://futureoflife.org/ai-safety-research/"><em>AI safety research grants</em></a><em>, which were funded by generous donations from Elon Musk and the Open Philanthropy Project.</em></p><p><em>Originally published at </em><a href="https://futureoflife.org/2018/09/17/making-ai-safe-in-an-unpredictable-world-an-interview-with-thomas-g-dietterich/"><em>futureoflife.org</em></a><em> on September 17, 2018.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e262a13967a6" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Risks Posed By Lethal Autonomous Weapons]]></title>
            <link>https://medium.com/@FLIxrisk/the-risks-posed-by-lethal-autonomous-weapons-b242e00284d0?source=rss-e33e2d2a809c------2</link>
            <guid isPermaLink="false">https://medium.com/p/b242e00284d0</guid>
            <category><![CDATA[united-nations]]></category>
            <category><![CDATA[ai-safety]]></category>
            <category><![CDATA[international-relations]]></category>
            <category><![CDATA[autonomous-weapons]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Future of Life]]></dc:creator>
            <pubDate>Tue, 04 Sep 2018 16:31:17 GMT</pubDate>
            <atom:updated>2018-09-14T20:30:01.409Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*phrm4uHaPSkpKm128IkGSw.jpeg" /></figure><p><em>By Ariel Conn</em></p><p><em>The following article was originally posted on </em><a href="https://metro.co.uk/2018/09/02/killer-robots-are-fast-becoming-a-reality-we-must-stop-this-from-happening-if-we-want-to-stop-a-global-ai-arms-race-7903717/"><em>Metro</em></a><em>.</em></p><p>Killer robots. It’s a phrase that’s both terrifying, but also one that most people think of as still in the realm of science fiction. Yet weapons built with artificial intelligence (AI) — weapons that could identify, target, and kill a person all on their own — are quickly moving from sci-fi to reality.</p><p>To date, no weapons exist that can specifically target people. But there are weapons that can track incoming missiles or locate enemy radar signals, and these weapons can autonomously strike these non-human threats without any person involved in the final decision. Experts predict that in just a few years, if not sooner, this technology will be advanced enough to use against people.</p><p>Over the last few years, delegates at the United Nations have debated whether to consider banning killer robots, more formally known as lethal autonomous weapons systems (LAWS). This week delegates met again to consider whether more meetings next year could lead to something more tangible — a political declaration or an outright ban.</p><p>Meanwhile, those who would actually be responsible for designing LAWS — the AI and robotics researchers and developers — have spent these years calling on the UN to negotiate a treaty banning LAWS. More specifically, nearly 4,000 AI and robotics researchers <a href="http://futureoflife.org/open-letter-autonomous-weapons/">called for a ban on LAWS in 2015</a>; in 2017, <a href="http://futureoflife.org/autonomous-weapons-open-letter-2017/">137 CEOs of AI companies asked the UN to ban LAWS</a>; and in 2018, 240 AI-related organizations and nearly 3,100 individuals took that call a step further and <a href="http://futureoflife.org/lethal-autonomous-weapons-pledge/">pledged not to be involved in LAWS development</a>.</p><p>And AI researchers have plenty of reasons for their consensus that the world should seek a ban on lethal autonomous weapons. Principle among these is that AI experts tend to recognize how dangerous and destabilizing these weapons could be.</p><p>The weapons could be hacked. The weapons could fall into the hands of “bad actors.” The weapons may not be as “smart” as we think and could unwittingly target innocent civilians. Because the materials necessary to build the weapons are cheap and easy to obtain, military powers could mass-produce these weapons, increasing the likelihood of proliferation and mass killings. The weapons could enable assassinations or, alternatively, they could become weapons of oppression, allowing dictators and warlords to subdue their people.</p><p>But perhaps the greatest risk posed by LAWS, is the potential to ignite a global AI arms race.</p><p>For now, governments insist they will ensure that testing, validation, and verification of these weapons is mandatory. However, these weapons are not only technologically novel, but also transformative; they have been described as the third revolution in warfare, following gun powder and nuclear weapons. LAWS have the potential to become the most powerful types of weapons the world has seen.</p><p>Varying degrees of autonomy already exist in weapon systems around the world, and levels of autonomy and advanced AI capabilities in weapons are increasing rapidly. If one country were to begin substantial development of a LAWS program — or even if the program is simply perceived as substantial by other countries — an AI arms race would likely be imminent.</p><p>During an arms race, countries and AI labs will feel increasing pressure to find shortcuts around safety precautions. Once that happens, every threat mentioned above becomes even more likely, if not inevitable.</p><p>As stated in the <a href="https://futureoflife.org/open-letter-autonomous-weapons/">Open Letter Against Lethal Autonomous Weapons</a>:</p><blockquote><em>The key question for humanity today is whether to start a global AI arms race or to prevent it from starting. If any major military power pushes ahead with AI weapon development, a global arms race is virtually inevitable, and the endpoint of this technological trajectory is obvious: autonomous weapons will become the Kalashnikovs of tomorrow. Unlike nuclear weapons, they require no costly or hard-to-obtain raw materials, so they will become ubiquitous and cheap for all significant military powers to mass-produce. It will only be a matter of time until they appear on the black market and in the hands of terrorists, dictators wishing to better control their populace, warlords wishing to perpetrate ethnic cleansing, etc. Autonomous weapons are ideal for tasks such as assassinations, destabilizing nations, subduing populations and selectively killing a particular ethnic group.</em></blockquote><p>Most countries here have expressed their strong desire to move from talking about this topic to reaching an outcome. There have been many calls from countries and groups of countries to negotiate a new treaty to either prohibit LAWS and/or affirm meaningful human control over the weapons. Some countries have suggested other measures such as a political declaration. But a few countries — especially Russia, the United States, South Korea, Israel, and Australia — are obfuscating the process, which could lead us closer to an arms race.</p><p>This is a threat we must prevent.</p><p><em>Originally published at </em><a href="https://futureoflife.org/2018/09/04/the-risks-posed-by-lethal-autonomous-weapons/"><em>futureoflife.org</em></a><em> on September 4, 2018.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=b242e00284d0" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>