<?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 Louis Rosenberg, PhD on Medium]]></title>
        <description><![CDATA[Stories by Louis Rosenberg, PhD on Medium]]></description>
        <link>https://medium.com/@louis_rosenberg?source=rss-46e426f3cb75------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*osMGQd6VU-ubnGdoOa6Llw.jpeg</url>
            <title>Stories by Louis Rosenberg, PhD on Medium</title>
            <link>https://medium.com/@louis_rosenberg?source=rss-46e426f3cb75------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Mon, 16 Nov 2026 06:02:14 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@louis_rosenberg/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[The Day after Superintelligence]]></title>
            <link>https://medium.com/predict/the-day-after-superintelligence-795888e5628c?source=rss-46e426f3cb75------2</link>
            <guid isPermaLink="false">https://medium.com/p/795888e5628c</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[social-media]]></category>
            <category><![CDATA[psychology]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[marketing]]></category>
            <dc:creator><![CDATA[Louis Rosenberg, PhD]]></dc:creator>
            <pubDate>Fri, 05 Dec 2025 18:45:32 GMT</pubDate>
            <atom:updated>2025-12-05T18:45:32.335Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*sAKeX_4O3o1h3uLJVK1sbQ.png" /><figcaption>The Day After Superintelligence</figcaption></figure><p><strong>Try this little experiment</strong>: Ask anyone you know if they’re concerned about AI and they will likely share a <a href="https://www.pewresearch.org/internet/2025/04/03/how-the-us-public-and-ai-experts-view-artificial-intelligence/">variety of fears</a>, from disruptions in the job market and the reality-bending impacts of deepfakes, to the extreme power being concentrated in a handful of large companies. These are all valid concerns, but they reveal <a href="https://bigthink.com/the-present/the-rise-of-ai-denialism/">signs of collective denial</a> regarding the impact that AI will have on our lives within the next handful of years.</p><p>The fact is, we humans will likely unleash AI systems that can outperform most us on most cognitive tasks, from planning and reasoning, to problem-solving, strategic thinking, and raw <a href="https://doi.org/10.1007/s11633-025-1546-4">creativity</a>. These systems will solve complex problems in a fraction of a second that might take the smartest human experts, days, weeks, or even years to work through. This terrifies me and it’s not just because of the impacts on the job market, the flood of deepfakes, and the excessive power concentration in Big Tech.</p><p>No, I am worried about the unexpected risks — the dangers that could emerge in the <a href="https://venturebeat.com/ai/ai-denial-is-becoming-an-enterprise-risk-why-dismissing-slop-obscures-real"><em>best-case scenarios</em></a> where the superintelligence is helpful, unbiased, and benevolent. Such an ASI will have many positive uses, but it could also be <a href="https://venturebeat.com/ai/have-we-reached-peak-human/">deeply demoralizing</a> to our identity as humans. After all, the world will feel different when each of us knows that a smarter, faster, more creative intelligence is available on our devices than between our own ears.</p><p>So ask yourself: how will people act in this new reality? Will we reflexively seek advice from our AI assistants as we navigate every little challenge we encounter in our live? Or worse, will we learn to <a href="https://www.intechopen.com/online-first/1212008">trust our AI assistants</a> more than we trust our own thoughts and instincts?</p><p>Wait — before you answer, you need to update your mental model:</p><p>Currently we engage AI through a <em>Socratic framework </em>that requires us to ask questions and get answers (like Captain Kirk did <a href="https://bigthink.com/the-present/danger-conversational-ai/">aboard the Enterprise</a> in 1966.) But that’s old-school thinking. We now are entering a new era in which AI assistants will be integrated into body-worn devices that are equipped with cameras and microphones, enabling the AI to see what you see, hear what you hear, and <a href="https://venturebeat.com/ai/enter-the-whisperverse-how-ai-voice-agents-will-guide-us-through-our-days/">whisper advice into your ears</a> without you needing to ask. This is an <em>Augmentation framework</em> where the AI is not waiting for your query but <a href="https://www.intechopen.com/online-first/1212008">proactively guiding you</a>.</p><p>In other words, our future will be filled with AI assistants that ride shotgun in our lives, <a href="https://venturebeat.com/ai/2024-will-be-the-year-of-augmented-mentality/"><em>influencing our experiences</em></a> at every turn. In this world, the risk is not that we reflexively ask AI for advice before using our own brains. No, the risk is that we won’t need to ask — the advice will just stream into our eyes and ears, shaping our actions, <a href="https://ieeexplore.ieee.org/document/10099167">driving our decisions</a>, and solving our problems before we’ve had a chance to think for ourselves.</p><p>I refer to this framework as <a href="https://www.intechopen.com/online-first/1212008">Augmented Mentality</a> and it will soon hit society at scale through AI-powered glasses, earbuds, and pendants. This is the future of mobile computing, and it is already driving an arms race between Meta, Google, Samsung, and Apple, as they position themselves to produce the <a href="https://www.intechopen.com/online-first/1212008">context-aware AI devices</a> that will replace handheld phones.</p><p>Imagine walking down the street. You notice a coworker heading towards you on the sidewalk. You can’t remember his name, but your AI assistant does. It detects your hesitation and <a href="https://bigthink.com/the-future/the-whisperverse-the-future-of-mobile-computing-is-an-ai-voice-inside-your-head/">whispers the coworker’s name</a> into your ears. The AI also recommends that you ask the coworker about his wife, who had surgery recently. The coworker appreciates the sentiment, then asks you about your recent promotion, likely at the advice of his own AI.</p><p>Is this human empowerment, or a loss of human agency?</p><p>It will certainly feel like a superpower to have <a href="https://www.google.com/books/edition/Our_Next_Reality/k3LdEAAAQBAJ">an AI in your ear</a> that always has your back, ensuring you never forget a name, always have witty things to say, and are instantly alerted when someone you’re talking to <a href="https://www.youtube.com/watch?v=IsE_Pas2OQU">is not being truthful</a>. On the other hand, everyone you meet will have their own AI muttering in their own ears. This will make us wonder who we’re <em>really</em> interacting with — the human standing in front of us, or the AI agent giving them guidance (see the 2021 short story <a href="https://08baef5b-6ceb-42d8-86e8-3107b583fa0b.filesusr.com/ugd/c7b69e_de242bcb09284c8789934bd192de168a.pdf">Carbon Dating</a> for fun examples).</p><p>Many experts believe that body-worn AI assistants will make us feel powerful, but that’s not the only way this could go. These same AI agents could make us <a href="https://venturebeat.com/ai/what-happens-the-day-after-superintelligence">feel less confident in ourselves</a> and less impactful in our lives. After all, human intelligence is a defining feature of humanity, a thing we take pride in as a species, and yet we could soon find ourselves deferring to AI assistants because we feel mentally outmatched. Is this empowerment — an AI that <a href="https://venturebeat.com/ai/what-happens-the-day-after-superintelligence"><em>botsplains</em></a> our world to us in real time?</p><p>I raise these concerns as someone who has spent my career building tools that <em>augment human abilities</em>. From my early work developing <a href="https://spectrum.ieee.org/history-of-augmented-reality">augmented reality</a> to my current work developing <a href="https://arxiv.org/abs/2401.15109">conversational agents</a> that make human teams smarter, my focus is using technology to enhance human abilities. Unfortunately, when it comes to superintelligence, there is a fine line between <em>augmenting our abilities</em> and <em>replacing them</em>. Unless we are thoughtful in how we deploy ASI, I fear we will cross that line.</p><p><a href="https://sites.google.com/view/louisrosenberg/bio">Louis Rosenberg</a> PhD is an early pioneer of virtual and augmented reality and longtime AI researcher. He founded Immersion Corp, Outland Research, and Unanimous AI, and was a tenured professor at California State University. His most recent book is <a href="https://www.google.com/books/edition/Our_Next_Reality/k3LdEAAAQBAJ">Our Next Reality</a> from Hachette.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=795888e5628c" width="1" height="1" alt=""><hr><p><a href="https://medium.com/predict/the-day-after-superintelligence-795888e5628c">The Day after Superintelligence</a> was originally published in <a href="https://medium.com/predict">Predict</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Ai agents and ‘The Game of Humans’]]></title>
            <link>https://medium.com/predict/ai-agents-and-the-game-of-humans-6bb7f59a71e8?source=rss-46e426f3cb75------2</link>
            <guid isPermaLink="false">https://medium.com/p/6bb7f59a71e8</guid>
            <category><![CDATA[social-media]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[marketing]]></category>
            <category><![CDATA[chatgpt]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Louis Rosenberg, PhD]]></dc:creator>
            <pubDate>Wed, 26 Feb 2025 02:26:15 GMT</pubDate>
            <atom:updated>2025-04-01T17:23:17.071Z</atom:updated>
            <content:encoded><![CDATA[<h3>Ai Agents and ‘The Game of Humans’</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SITZzG3v58cTiu5yM5PJkA.jpeg" /></figure><p>When I was a kid there were four <a href="https://venturebeat.com/ai/why-context-aware-ai-agents-will-give-us-superpowers-in-2025/">AI agents</a> in my life. Their names were Inky, Blinky, Pinky, and Clyde and they tried their best to hunt me down. This was 1980 and the agents were the four colorful ghosts in the iconic arcade game Pac-Man.</p><p>By today’s standards they weren’t particularly smart, and yet they seemed to pursue me with cunning and intent. This was decades before <a href="https://venturebeat.com/ai/googles-new-neural-net-architecture-separates-memory-components-to-control-exploding-costs/">neural networks</a> were used in video games, so their behaviors were controlled by simple algorithms called <em>heuristics</em> that dictate how they would chase me around the maze.</p><p>Most people don’t realize this, but the four ghosts were designed with different “<a href="https://pacman.fandom.com/wiki/Maze_Ghost_AI_Behaviors">personalities</a>.” Good players can observe their actions and learn to predict their behaviors. For example, the red ghost (Blinky) was programmed with a “pursuer” personality that charges directly towards you. The pink ghost (Pinky) on the other hand, was given an “ambusher” personality that predicts where you’re going and tries to get there first. As a result, if you rush directly at Pinky, you can use her personality against her, causing her to actually turn away from you.</p><p>I reminisce because in 1980 a skilled human could observe these AI agents, decode their unique personalities, and use those insights to outsmart them. Now, 45 years later, the tides are about to turn. Like it or not, AI agents will soon be deployed that are tasked with decoding <em>your personality</em> so they can use those insights to <a href="https://arxiv.org/abs/2306.11748">optimally influence</a> you.</p><p>In other words, we are all about to become unwitting players in “The Game of Humans” and it will be the AI agents trying to earn the high score. I mean this literally — most AI systems are designed to maximize a “<a href="https://en.wikipedia.org/wiki/Reinforcement_learning">reward function</a>” that earns points for achieving objectives. This allows AI systems to quickly find optimal solutions. Unfortunately, without regulatory protections, we humans will likely become <em>the objective </em>that AI agents are<a href="https://venturebeat.com/ai/enter-the-whisperverse-how-ai-voice-agents-will-guide-us-through-our-days/"> tasked with optimizing</a>.</p><p>I am most concerned about the <a href="https://arxiv.org/pdf/2202.03164">conversational agents</a> that will engage us in friendly dialog throughout our daily lives. They will speak to us through photorealistic avatars on our PCs and phones, and soon, through<a href="https://www.technologyreview.com/2025/02/05/1110983/whats-next-for-smart-glasses/"> AI-powered glasses</a> that guide us through our days. Unless there are clear restrictions, these agents will be designed to conversationally probe us for information so they can characterize our temperaments, tendencies, personalities and desires, and use those traits to <a href="https://www.researchgate.net/publication/368492998_The_Metaverse_and_Conversational_AI_as_a_Threat_Vector_for_Targeted_Influence">maximize their persuasive impact</a> when working to sell us products, pitch us services, or convince us to believe misinformation.</p><p>This is called <a href="https://www.researchgate.net/publication/369355910_The_Manipulation_Problem_Conversational_AI_as_a_Threat_to_Epistemic_Agency">the AI Manipulation Problem</a> and I’ve been warning regulators about the risk<a href="https://futurism.com/artificial-intelligence-isnt-intelligence-manipulating-humanity"> since 2016</a>. Thus far, policymakers have not taken decisive action, viewing the threat as too far in the future. But now, with the release of Deepseek-R1, the final barrier to widespread deployment of AI agents, <em>the cost of real-time processing</em>, has been greatly reduced. Before this year is out, AI agents will become a<a href="https://www.iiis.org/DOI2023/HA408FU/"> new form of targeted media</a> that is so<a href="https://ieeexplore.ieee.org/document/9965661"> interactive and adaptive</a>, it can optimize its ability to influence our thoughts, guide our feelings, and drive our behaviors.</p><p>Of course, human salespeople are interactive and adaptive too. They engage us in friendly dialog to size us up, quickly finding the buttons they can press to sway us.<a href="https://venturebeat.com/ai/agents-of-manipulation-the-real-ai-risk/"> AI agents</a> will make them look like amateurs, able to draw information out of us with such finesse, it would intimidate a seasoned therapist. And they will use these insights to adjust their conversational tactics in real-time, working to<a href="https://futurism.com/artificial-intelligence-isnt-intelligence-manipulating-humanity"> persuade us</a> more effectively than any used car salesman.</p><p>These will be<a href="https://www.iiis.org/DOI2023/HA408FU/"> asymmetric encounters</a> in which the artificial agent has the upper hand. After all, when you engage a human who is trying to influence you, you can usually sense their motives and honesty. It will not be a fair fight with AI agents. They will be able to size you up with superhuman skill, but you won’t be able to size them up at all. That’s because they will look, sound, and act so human, we will<a href="https://pubmed.ncbi.nlm.nih.gov/35165187/"> unconsciously trust them</a> when they smile with empathy and understanding, forgetting that their facial affect is just a simulated façade.</p><p>In addition, their voice, vocabulary, speaking style, age, gender, race and facial features are likely to be customized for each of us personally to<a href="https://arxiv.org/abs/2306.11748"> maximize our receptiveness</a>. And unlike human salespeople who need to size up each customer from scratch, these virtual entities could have access to stored data about our backgrounds and interests. They could then use this personal data to quickly<a href="https://bigthink.com/the-future/the-whisperverse-the-future-of-mobile-computing-is-an-ai-voice-inside-your-head/"> earn your trust</a>, asking you about your kids, your job, or maybe your beloved Yankees, easing you into subconsciously letting down your guard.</p><p>To educate policymakers on the risk of AI-powered manipulation, I helped in the making of an award-winning short film entitled<a href="https://www.privacylost.org/"> <strong>Privacy Lost</strong></a> that was produced by the Responsible Metaverse Alliance, Minderoo, and the XR Guild. The quick<a href="https://www.youtube.com/watch?v=IsE_Pas2OQU"> 3-minute narrative</a> depicts a young family eating in a restaurant while wearing AR glasses. Instead of human servers, avatars take each diner’s orders, using the power of AI to upsell them in personalized ways. The film was considered sci-fi when released in 2023, and yet only two years later, Big Tech is engaged in an all-out <a href="https://www.technologyreview.com/2025/02/05/1110983/whats-next-for-smart-glasses/">arms race to make AI-powered eyewear</a> that could easily be used in these ways.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3fIlMqM1GBdQBPCXM8dFBQ.jpeg" /><figcaption>Still Image from Award Winning Short Film <a href="https://www.youtube.com/watch?v=IsE_Pas2OQU">PRIVACY LOST</a>.</figcaption></figure><p>In addition, we need to consider the psychological impact that will occur when we humans start to believe that the AI agents giving us advice are <a href="https://venturebeat.com/ai/have-we-reached-peak-human/"><em>smarter than us</em></a> on nearly every front. When AI achieves a perceived state of <em>“Cognitive Supremacy” </em>with respect to the average human, it will likely cause us to blindly <a href="https://venturebeat.com/ai/enter-the-whisperverse-how-ai-voice-agents-will-guide-us-through-our-days/">accept its guidance</a> rather than using our own critical thinking. This <em>deference</em> to a perceived superior intelligence (whether truly superior or not) will make agent manipulation that much easier to deploy.</p><p>I am not a fan of over-regulation, but we need smart, narrow restrictions on AI to avoid<a href="https://venturebeat.com/ai/why-generative-ai-is-more-dangerous-than-you-think/"> <em>superhuman manipulation</em></a> by conversational agents. Without protections, these agents will convince us to buy things we don’t need, believe things that are untrue, and accept things that are not in our best interest. It’s easy to tell yourself you won’t be susceptible, but with AI optimizing every word they say to us, it is likely we will all be outmatched.</p><p>One solution is to ban AI agents from establishing<a href="https://library.xrguild.org/science-and-ethics-council-topics/conversational-agents"> <em>feedback loops</em></a> in which they optimize their persuasiveness by analyzing our reactions and repeatedly adjusting their tactics. In addition, AI agents should be required to inform you of their objectives. If their goal is to convince you to buy a car, vote for a politician, or pressure your family doctor for a new medication, those objectives should be stated up front. And finally, AI agents should not have access to personal data about your background, interests, or personality if such data can be<a href="https://venturebeat.com/ai/the-profound-danger-of-conversational-ai/"> used to sway you</a>.</p><p>In today’s world, targeted influence is an overwhelming problem, and it is mostly deployed as buckshot fired in your general direction. Interactive AI agents will turn targeted influence into<a href="https://arxiv.org/abs/2306.11748"> heat-seeking missiles</a> that find the best path into each of us. If we don’t protect against this risk, I fear we could all lose <em>the game of humans</em>.</p><p>This originally <a href="https://venturebeat.com/ai/like-it-or-not-ai-is-learning-how-to-influence-you/">appeared in VentureBeat (2/15/25)</a></p><p><a href="https://sites.google.com/view/louisrosenberg/bio">Louis Rosenberg, PhD</a> is a computer scientist and entrepreneur. He is known for pioneering<a href="https://spectrum.ieee.org/history-of-augmented-reality"> augmented reality</a> at Air Force Research Laboratory and for founding <a href="https://unanimous.ai/">Unanimous AI</a> and Immersion Corporation. He earned his PhD from Stanford, was a professor at California State University, and has been awarded over 300 patents for his work in AI, XR, and HCI.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=6bb7f59a71e8" width="1" height="1" alt=""><hr><p><a href="https://medium.com/predict/ai-agents-and-the-game-of-humans-6bb7f59a71e8">Ai agents and ‘The Game of Humans’</a> was originally published in <a href="https://medium.com/predict">Predict</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Big Tech aims to make us Superhuman]]></title>
            <link>https://medium.com/predict/big-tech-aims-to-make-us-superhuman-a8265688c155?source=rss-46e426f3cb75------2</link>
            <guid isPermaLink="false">https://medium.com/p/a8265688c155</guid>
            <category><![CDATA[augmented-reality]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[marketing]]></category>
            <category><![CDATA[social-media]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Louis Rosenberg, PhD]]></dc:creator>
            <pubDate>Sun, 12 Jan 2025 22:24:22 GMT</pubDate>
            <atom:updated>2025-04-01T17:26:19.665Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*25g1z2N3cPT34z9DWvZGFw.jpeg" /></figure><p><strong>2025 will be the year </strong>that Big Tech transitions from selling us more and more powerful <em>tools </em>to selling us more and more powerful <em>abilities</em>. The difference between a tool and an ability is subtle yet profound. We use tools as <em>external artifacts</em> that help us overcome our organic limitations. From cars and planes to phones and computers, tools greatly expand what we can accomplish as individuals, in large teams, and as vast civilizations.</p><p><strong>Abilities are different.</strong> We experience abilities in the first person as <em>self-embodied </em>capabilities that <em>feel internal</em> and <em>instantly accessible</em> to our conscious minds. For example, language and mathematics are human created <em>technologies</em> that we load into our brains and carry around with us throughout our lives, expanding our abilities to think, create, and collaborate. They are<a href="https://bigthink.com/the-future/ar-will-make-us-superhumans/"> </a><a href="https://venturebeat.com/datadecisionmakers/augmented-reality-superhuman-abilities-and-the-future-of-medicine/">superpowers</a> that feel so inherent to our existence, we rarely think of them as technologies at all. Fortunately, we don’t need to buy a service plan.</p><p>The next wave of superpowers, however, <strong>will not be free</strong>. But just like our abilities to think verbally and numerically, we will experience these powers as self-embodied capabilities that we carry around with us throughout our lives. I refer to this new technological discipline as<a href="https://venturebeat.com/ai/2024-will-be-the-year-of-augmented-mentality/"> <em>augmented mentality</em></a> and it will emerge from the convergence of artificial intelligence, conversational computing, and augmented reality. And in 2025 it will kick off an arms race among the largest companies in the world to sell us<a href="https://bigthink.com/the-future/ar-will-make-us-superhumans/"> superhuman abilities</a>.</p><p>These new superpowers will be unleashed by<a href="https://www.researchgate.net/publication/367183000_The_Metaverse_The_Ultimate_Tool_of_Persuasion"> context-aware AI agents</a> that are loaded into<a href="https://venturebeat.com/datadecisionmakers/augmented-reality-superhuman-abilities-and-the-future-of-medicine/"> </a><a href="https://venturebeat.com/datadecisionmakers/future-augmented-reality-will-inherit-the-earth/">body worn devices</a> (like AI glasses) that travel with us throughout our lives, seeing what we see, hearing what we hear, experiencing what we experience, and providing us with enhanced abilities to perceive and interpret our world. In fact, by 2030, I predict that a majority of us will live our lives with the aid of context-aware AI agents that bring<a href="https://medium.com/predict/augmented-reality-will-give-us-superpowers-e0a2f4a8d777"> digital superpowers</a> into our normal daily experiences.</p><p><strong>How will our Superhuman Future unfold?</strong></p><p>First and foremost, we will <a href="https://venturebeat.com/ai/enter-the-whisperverse-how-ai-voice-agents-will-guide-us-through-our-days/">whisper</a> to these intelligent agents, and they will <a href="https://techcrunch.com/2022/01/12/the-metaverse-will-be-filled-with-elves">whisper back</a>, acting like an omniscient alter ego that gives us context-aware recommendations, knowledge, guidance, advice, <a href="https://patents.google.com/patent/US7577522B2/en">spatial reminders</a>, directional cues, <a href="https://medium.com/@history-of-immersive-computing/history-of-haptic-computing-538a84bb4303">haptic nudges</a>, and other verbal and perceptual content that will coach us through our days and educate us about our world.</p><p><strong>Consider this simple scenario</strong>: You are walking downtown and spot a store across the street. You wonder, <em>what time does it open?</em> So, you grab your phone and type (or say) the name of the store. You quickly find the hours on a website and maybe review other info about the store as well. That is the basic <em>tool-use computing </em>model prevalent today.</p><p>Now, let’s look at how Big Tech will transition to an <em>ability computing</em> model:</p><p><strong>Phase 1</strong>: You are wearing AI-powered glasses that can see what you see, hear what you hear, and process your surroundings through a multimodal large language model. Now when you spot that store across the street, you simply whisper to yourself — <em>“I wonder when it opens?” </em>and a voice will instantly ring back into your ears “10:30 AM.”</p><p>I know this is a subtle shift from asking your phone to look up the name of a store, but it will feel profound. The reason is that the context-aware AI agent will <em>share your reality</em>. It’s not just tracking your location like GPS, it is seeing what you see, hearing what you hear, and paying attention to what you are paying attention to. This will make it feel far less like a tool, and far more like an internal ability that is linked to your first-person reality.</p><p>And when we are asked a question by the AI-powered alter ego in our ears, we will often answer by just <a href="https://patents.google.com/patent/US7489979B2">nodding our heads to affirm</a> (detected by sensors in the glasses) or shaking our heads to reject. It will feel so natural and seamless, we might not even consciously realize we replied.</p><p><strong>Phase 2</strong>: By 2030, we will not need to whisper to the AI agents traveling with us through our lives. Instead, you will be able to simply mouth the words, and the AI will know what you are saying by reading your lips and detecting activation signals from your muscles. I am confident that <em>“mouthing”</em> will be deployed because it’s more private, more resilient to noisy spaces, and most importantly, it will feel more personal, internal, and self-embodied.</p><p><strong>Phase 3</strong>: By 2035, you may not even need to mouth the words. That’s because the AI will learn to interpret the signals in our muscles with such subtlety and precision, we will simply need <em>to think about mouthing the words</em> to convey our intent. You will be able to focus your attention on any item or activity in your world and <em>think</em> <em>something, </em>and useful information will ring back from your AI glasses like an<a href="https://bigthink.com/the-future/the-whisperverse-the-future-of-mobile-computing-is-an-ai-voice-inside-your-head/"> all-knowing voice</a> in your head.</p><p>Of course, the capabilities will go far beyond just wondering about things around you. That’s because the onboard AI that shares your first-person reality will learn to anticipate the information you desire before you even ask for it. For example, when a coworker approaches from down the hall, and you can’t quite remember his name, the AI will sense your unease, and a voice will ring — <em>“Gregg from engineering.”</em></p><p>Or when you pick up a can of soup in a store and are curious about the carbs, or wonder if it’s cheaper at Walmart, the answers will just ring in your ears or appear visually. It will even give you superhuman abilities to assess the emotions on other people’s faces, predict their moods, goals, or intentions, and coach you during real-time conversations to make you more compelling, appealing or persuasive. See fun<a href="https://www.youtube.com/watch?v=IsE_Pas2OQU"> video example</a>:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FIsE_Pas2OQU%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DIsE_Pas2OQU&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FIsE_Pas2OQU%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/95e67ff265ca6fea45791c6e899cdb16/href">https://medium.com/media/95e67ff265ca6fea45791c6e899cdb16/href</a></iframe><p>I know some people will be skeptical about the level of adoption I predict above and the rapid time frame, but I don’t make these claims lightly. I’ve spent most of my career working on <a href="https://medium.com/predict/the-metaverse-of-the-1990s-d7d6ac9c3de2">VR</a>, <a href="https://commons.wikimedia.org/wiki/File:Virtual_Fixtures_Project_(1991-1994)_animated_example.gif">AR</a>, and <a href="https://medium.com/predict/how-to-build-a-collective-mind-that-speaks-for-humanity-in-real-time-eade7f60c292">AI</a> technologies that<a href="https://spectrum.ieee.org/history-of-augmented-reality"> augment and expand human abilities</a>, and I can say without question, the mobile computing market is about to run in this direction in a very big way.</p><p>Over the last 12 months, two of the most influential and innovative companies in the world, Meta and Google, revealed their intentions to give us self-embodied superpowers. Meta made the first big move by adding a context-aware AI to their Ray-Ban glasses and by showing off their Orion mixed reality prototype that adds impressive visual capabilities. Meta is now very well positioned to leverage their big investments in AI and XR and become a major player in the mobile computing market and they will likely do it by selling us superpowers we can’t resist.</p><p>Not to be outdone, Google recently<a href="https://blog.google/products/android/android-xr/"> announced Android XR</a>, a new AI-powered operating system for<a href="https://link.springer.com/chapter/10.1007/978-3-030-89906-6_1"> augmenting our world</a> with seamless context-aware content. They also announced a partnership with Samsung to bring new glasses and headsets to market. With over 70% market-share for mobile operating systems and an increasingly strong AI presence with Gemini, I believe that Google is well-positioned to be the leading provider of technology-enabled human superpowers within the next few years.</p><p><strong>Of course, we need to consider the risks.</strong></p><p>To quote the famous<a href="https://www.marvel.com/comics/issue/16926/amazing_fantasy_1962_15"> 1962 Spiderman comic</a>, “with great power comes great responsibility.” This wisdom is <em>literally </em>about superpowers. The difference is that the <em>great responsibility</em> will not fall on the consumers who purchase these techno-powers, but on the companies that provide them and the regulators that oversee them.</p><p>After all, when wearing AI-powered AR eyewear, each of us could find ourselves in a <a href="https://www.amazon.com/Our-Next-Reality-AI-powered-Metaverse/dp/1399812246">new reality</a> where technologies<a href="https://dl.acm.org/doi/full/10.1145/3614426"> controlled by third parties</a> can selectively alter what we see and hear, while AI-powered voices whisper in our ears with advice, information, and guidance. While the intentions are magical, the<a href="https://www.researchgate.net/publication/368492998_The_Metaverse_and_Conversational_AI_as_a_Threat_Vector_for_Targeted_Influence"> potential for abuse</a> is just as profound.</p><p><strong>To avoid dystopian outcomes</strong>, my primary recommendation to users and manufacturers is to<a href="https://www.researchgate.net/publication/362541437_Regulating_the_Metaverse_a_Blueprint_for_the_Future"> adopt a subscription business model</a>. If the arms race for selling superpowers is driven by which company can provide the most amazing new abilities for a reasonable monthly fee — we will all benefit. If instead, the business model becomes a competition to monetize superpowers by delivering the most effective targeted influence into our eyes and ears throughout our daily lives, consumers could<a href="https://github.com/GenerativeAIandHCI/GenerativeAIandHCI.github.io/blob/main/papers/2023/Conversational%20AI%20and%20the%20Threat%20to%20Epistemic%20Agency%20-%20CHI%202023%20.pdf"> easily be manipulated</a> with precision and pervasiveness we have never before faced.</p><p>Ultimately, these superpowers won’t feel optional. After all, not having them could put us at a cognitive disadvantage. It is now up to the industry and regulators to ensure that we roll out these new abilities in a way that is not <a href="https://venturebeat.com/ai/the-hidden-dangers-of-generative-advertising/">intrusive</a>, <a href="https://venturebeat.com/ai/agents-of-manipulation-the-real-ai-risk/">manipulative</a>, or <a href="https://venturebeat.com/ai/the-profound-danger-of-conversational-ai/">dangerous</a>. I believe this can be a truly magical direction for computing, but the need for careful planning and oversight cannot be understated.</p><p><a href="https://sites.google.com/view/louisrosenberg/bio"><em>Louis Rosenberg</em></a><em> PhD is an early pioneer of virtual and augmented reality and a longtime AI researcher. He founded Immersion Corp, Outland Research, and Unanimous AI, and authored the recent book </em><a href="https://www.amazon.com/Our-Next-Reality-AI-powered-Metaverse/dp/1399812246"><em>Our Next Reality</em></a><em>.</em></p><p><em>A version of this piece originally appeared in </em><strong><em>VentureBeat</em></strong><em> (</em><a href="https://venturebeat.com/ai/why-context-aware-ai-agents-will-give-us-superpowers-in-2025/"><em>VB 1.5.25</em></a><em>)</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a8265688c155" width="1" height="1" alt=""><hr><p><a href="https://medium.com/predict/big-tech-aims-to-make-us-superhuman-a8265688c155">Big Tech aims to make us Superhuman</a> was originally published in <a href="https://medium.com/predict">Predict</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Have we reached Peak Human?]]></title>
            <link>https://medium.com/predict/have-we-reached-peak-human-d6b1c2d9c58a?source=rss-46e426f3cb75------2</link>
            <guid isPermaLink="false">https://medium.com/p/d6b1c2d9c58a</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[future]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[marketing]]></category>
            <category><![CDATA[large-language-models]]></category>
            <dc:creator><![CDATA[Louis Rosenberg, PhD]]></dc:creator>
            <pubDate>Sun, 22 Sep 2024 01:19:54 GMT</pubDate>
            <atom:updated>2024-09-28T12:05:48.574Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Pn6RYCxEJvA5xUhI9EmgUw.png" /><figcaption>Have we reached Peak Human?</figcaption></figure><p>I’m not talking about <em>peak population</em> or <em>peak lifespan </em>or <em>peak performance</em>. When I refer to <strong>Peak Human</strong>, I’m talking about whether we have reached the point where AI systems can <em>outthink the majority of individual humans</em>, thus defining the moment in time when we “peaked” as an intellectual force on planet earth.</p><p>After all, once we pass this milestone, we will steadily <em>lose our cognitive edge</em> until AI systems can outthink all individual humans — even the most brilliant among us. At that later point, we will say that AI has achieved the more dangerous milestone of Superintelligence. And while most people focus on the second milestone, I believe we should track the first because it could happen within months, not years.</p><p>Until recently, the average human could easily outperform even the most powerful AI systems when it comes to basic reasoning. There are many ways to measure reasoning, none-of-which are considered the gold standard, but the best known is the classic IQ test. A journalist named Maxim Lott has been testing Large Language Models on a standardized Mensa IQ test. Last week, for the first time, an AI model significantly beat the median human IQ of 100. The model that crossed the peak of the bell curve was Open AI’s new “o1” system — it<a href="https://www.maximumtruth.org/p/massive-breakthrough-in-ai-intelligence"> reportedly scored 120 IQ</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/722/1*OJH0hVavaq3EQUCYGhjjjw.png" /></figure><p>So, does this mean AI has exceeded the reasoning ability of most humans?</p><p><strong>Not so fast</strong>. It is not valid to administer a standard IQ test to an AI system because the data it trained on likely included the tests and answers. To address this, Lott had a custom IQ test created that does not appear anywhere online and therefore is not in the training data. He gave that “offline test” to Open AI’s new “o1” model and it<a href="https://trackingai.org/IQ"> scored an IQ of 95</a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/680/1*DlX1e5Y06qCy9MYc_VS1aQ.jpeg" /><figcaption>Open AI’s o1 model scores 95 when given a test that it was NOT trained on.</figcaption></figure><p>This is still an impressive result. That score beats 37% of adults on the reasoning tasks. It also represents a rapid increase, as Open AI’s prior model GPT 4, released just last year, was outperformed by 98% of adults on the same test. At this rate of progress, it is very likely that an AI model will be able to beat 50% of adult humans on standard IQ tests this year.</p><p>I know this upsets people and there is “gut reaction” to push back on this view and point out that despite the performance of Open AI’s new o1 model on logic and reasoning tests, <em>it still makes dumb mistakes</em> at times. That’s true, but so do most humans. Probably all humans. I also hear people say that LLMs use “pattern matching” and “memorization” to solve many problems. Again, very true, but so do humans.</p><p><strong>Does this mean we will reach Peak Human in 2024? </strong><em>Yes and no.</em></p><p>First, <strong>I predict yes,</strong> at least one foundational AI model will be released in 2024 that can outthink more than 50% of adult humans on pure reasoning tasks. From this perspective, we will exceed my definition for Peak Human and will be on a downward path towards the rapidly approaching day when an AI is released that can outperform all individual humans, period.</p><p>Second, <strong>I say no so fast</strong>, because <strong>we humans have another trick </strong>up our sleeves. It’s called collective intelligence, and it relates to the fact that human groups can be smarter than individuals. And we humans have a lot of individuals — over 8 billion at the moment.</p><p>I bring this up because my personal focus as an AI researcher over the last decade has been the use of AI to connect groups of humans together into real-time systems that amplify our collective intelligence to superhuman levels. I call this goal<a href="https://venturebeat.com/ai/the-promise-of-collective-superintelligence/"> Collective Superintelligence</a> and I believe it is a viable pathway for keeping humanity cognitively competitive even after AI systems can outperform the reasoning ability of every individual among us. I like to think of this as “Peak Humanity,” and I am confident we can push it to intelligence levels that will surprise us all.</p><p>Back in 2019, my research team at<a href="https://unanimous.ai/"> Unanimous AI</a> conducted our first experiments in which we enabled groups of people to take IQ tests together by forming real-time systems mediated by Ai algorithms. This first-generation technology called Swarm AI enabled small groups of 6 to 10 randomly selected participants (who averaged 100 IQ) to amplify their collective performance to a collective IQ score of 114 when deliberating as an AI facilitated system (<a href="https://www.researchgate.net/publication/336221485_Swarm_Intelligence_Amplifies_the_IQ_of_Collaborating_Teams">Willcox &amp; Rosenberg</a>). This was a good start, but not within striking distance of Collective Superintelligence.</p><p>More recently we unveiled a new technology called<a href="https://www.youtube.com/watch?v=j_YfsD8s9-E"> Conversational Swarm Intelligence</a> (CSI). It enables large groups (up to 400 people) to hold real-time conversational deliberations that amplify the group’s collective intelligence. In collaboration with Carnegie Mellon University, we conducted a 2024 study in which groups of 35 randomly selected people were tasked with taking IQ test questions together in real-time as AI-facilitated “<a href="https://www.thinkscape.ai/thinkscape-technology">conversational swarms</a>” using a new CSI-powered online collaboration environment to called <a href="https://www.thinkscape.ai/">Thinkscape</a>. When taking the test using this online CSI platform, the<a href="https://www.scitepress.org/PublicationsDetail.aspx?ID=4yQ7CAluLXk=&amp;t=1"> groups averaged IQ scores of 128</a> (the 97th percentile). This is a strong result, but I believe we are just scratching the surface of how smart humans can become when we use AI to connect human groups together in larger and larger numbers.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/703/1*6VhTjIZsvscWMA1tmNTPww.jpeg" /><figcaption>Conversational Swarm of 35 people beats OPEN AI’s new 01 model on IQ tests</figcaption></figure><p>I am passionate about pursuing<a href="https://arxiv.org/abs/2311.00728"> Collective Superintelligence</a> because it has the potential to greatly amplify humanity’s cognitive abilities, and unlike a digital superintelligence it is inherently instilled with human values, morals, sensibilities, and interests. Of course, this begs the question — <em>how long can we stay ahead of the purely digital AI systems? </em>That depends on whether AI continues to advance at an accelerating pace or if we hit a plateau. Either way, amplifying our collective intelligence might help us maintain our edge long enough to figure out how to protect ourselves from being outmatched.</p><p>When I raise the issue of Peak Human, many people point out that human intelligence <strong>is far more than just the reasoning measured by IQ tests</strong>. I fully agree with this, but when we look deeper at some of the most “human” of all qualities — like creativity and artistry, we see evidence that AI systems are catching up with us just as quickly. It was only a few years ago that virtually all artwork created on planet earth was crafted by us humans. A<a href="https://journal.everypixel.com/ai-image-statistics"> recent analysis</a> estimates that Generative AI is producing 15 billion images per year and that rate is accelerating.</p><p>Even more surprising, a<a href="https://www.nature.com/articles/s41598-023-40858-3"> study published just last week</a> showed that AI Chatbots can outperform humans on creativity tests. To quote the paper, <em>“the results suggest that AI has reached at least the same level, or even surpassed, the average human’s ability to generate ideas in the most typical test of creative thinking (</em><a href="https://en.wikipedia.org/wiki/Guilford%27s_Alternate_Uses"><em>AUT</em></a><em>).”</em> I’m not sure I fully believe this result, but it’s just a matter of time before it holds true.</p><p>Whether we like it or not, our evolutionary position as the smartest and most creative brains on planet earth is likely to be challenged in the very near future. We can debate whether this will be a net positive or a net negative for humanity, but either way — we need to be doing more to<a href="https://venturebeat.com/ai/agents-of-manipulation-the-real-ai-risk"> protect ourselves</a> from being outmatched.</p><p><a href="https://sites.google.com/view/louisrosenberg/bio"><strong><em>Louis Rosenberg, PhD</em></strong></a><em> is a computer scientist and engineer. He earned his doctorate from Stanford, was a researcher NASA and the US Air Force, and founded a number of companies including Immersion Corp (IMMR: Nasdaq), Outland Research, and Unanimous AI. His latest book,</em><a href="https://www.amazon.com/Our-Next-Reality-Preparing-AI-powered/dp/1399812246/ref=sr_1_1"><em> Our Next Reality</em></a><em>, explores the impact of AI and spatial computing on humanity.</em></p><p>This piece originally appeared in <a href="https://venturebeat.com/ai/have-we-reached-peak-human/">VentureBeat (9/18/2024)</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d6b1c2d9c58a" width="1" height="1" alt=""><hr><p><a href="https://medium.com/predict/have-we-reached-peak-human-d6b1c2d9c58a">Have we reached Peak Human?</a> was originally published in <a href="https://medium.com/predict">Predict</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[How to build a Collective Mind that speaks for humanity in real-time.]]></title>
            <link>https://medium.com/predict/how-to-build-a-collective-mind-that-speaks-for-humanity-in-real-time-eade7f60c292?source=rss-46e426f3cb75------2</link>
            <guid isPermaLink="false">https://medium.com/p/eade7f60c292</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[science]]></category>
            <category><![CDATA[marketing]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[social-media]]></category>
            <dc:creator><![CDATA[Louis Rosenberg, PhD]]></dc:creator>
            <pubDate>Sun, 07 Jul 2024 00:33:59 GMT</pubDate>
            <atom:updated>2024-07-12T13:50:55.295Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*Xxf9XcttwnwFORaK4GUYzg.png" /><figcaption>How to build a global hive mind using Swarm Intelligence and Generative AI</figcaption></figure><p>In the world of science fiction, the notion that humanity will one day connect our brains together into a global “hive mind” dates back to a 1930 novel entitled “Last and First Men” by<a href="https://en.wikipedia.org/wiki/Olaf_Stapledon"> Olaf Stapledon</a>. It recounts a fictional “future history” in which humanity evolves biologically into a species that can link telepathically into collective minds of extreme intelligence.</p><p>In the real world this pursuit is called<a href="https://arxiv.org/abs/2401.15109"> Collective Superintelligence</a> and it will not require <em>telepathy</em> or other fictional devices. Instead, it will use Generative AI technologies to connect large human groups into real-time deliberative systems, enabling us to<a href="https://www.youtube.com/watch?v=Eu-RyZt_Uas"> solve difficult problems</a> by harnessing our combined knowledge, wisdom, and insight in powerful new ways.</p><p>This has been my personal focus as an AI researcher for the last decade and I believe it has the potential to produce<a href="https://ieeexplore.ieee.org/document/10452485?source=AUTHORALERT&amp;dld=eWFob28uY29t"> superintelligent systems</a> that maintain human values, morals, and interests at the core of every decision. Of course, for many people the idea of large human groups <em>thinking together</em> in real-time systems seems unnatural, but Mother Nature would disagree.</p><figure><img alt="Dr Louis Rosenberg talking about Collective Superintelligence at TED in 2017" src="https://cdn-images-1.medium.com/max/1024/1*AyAqCK2LOqMmmiWhYOHCQQ.jpeg" /><figcaption>Rosenberg talking about biological hive minds in 2017 (<a href="https://www.youtube.com/watch?v=Eu-RyZt_Uas">video TEDx Kansas City</a>)</figcaption></figure><p>In fact, many social species have evolved naturally in this direction, developing the ability to make rapid real-time decisions in large groups that greatly exceed the brainpower of individual members. Biologists call this<a href="https://venturebeat.com/business/how-swarm-intelligence-could-save-us-from-the-dangers-of-ai/"> Swarm Intelligence</a> and it enables schools of fish, swarms of bees, and flocks of birds to quickly make life-and-death decisions at intelligence levels that far exceed the mental capacity of their individual minds.</p><p>One remarkable form of Swarm Intelligence that has inspired my own<a href="https://link.springer.com/chapter/10.1007/978-3-030-29516-5_79"> academic research</a> over the last decade is the fish school. Although it looks simple on the surface, the inner dynamics enable thousands of members to make complex decisions with no individual in charge. In fact, fish schools can make good decisions even when no single member has sufficient information to address the problem at hand. Consider the image below:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*0WJZPVjbvNG4Wf3tCKoUPw.jpeg" /><figcaption>A fish school is a large organization with thousands of members and difficult problems to solve</figcaption></figure><p>The school above faces a hypothetical problem of extreme importance — three predators approach from three different directions. At the moment shown, no individual is aware of all three threats. In fact, most are not aware of any threats. There is a small group of fish in the upper left that is aware of the first predator, a small group in the lower left that is aware of the second predator, and a small group in the upper right that is aware of the third predator. Most of the other fish are unaware of the dangers.</p><p>So, how can a large group, in which all members have <strong>limited information</strong> (and no members possess all the information) solve this life and death problem and quickly move in the right direction? The first thing to know is that fish have a special organ on the sides of their body called the<a href="https://en.wikipedia.org/wiki/Lateral_line"> lateral line</a> that allows them to track the speed and direction of neighboring fish based on pressure and vibrations in the water around them. Using this organ, they can perceive the intent of their neighbors (i.e., the speed and direction those fish believe the school should go). This communication is bidirectional so we can think of it as a<a href="https://www.pnas.org/doi/epdf/10.1073/pnas.1007102108"> <em>local deliberation</em></a><em> </em>in which small groups decide the motion of their portion of the school.</p><p>This is interesting but does not explain how global decisions are made. After all, the group on the right that sees a predator approaching, likely decides the school should move left. At the same time, the two groups on the left likely decide the school should move right. And groups in the middle, which have no knowledge of the predators, are likely to keep moving in the direction they were already going. So, how does this get resolved into a single rapid decision that evades the urgent threats?</p><p>The magic happens because every fish in the school “deliberates” with a different group of nearby neighbors. This means there are many “overlapping conversations” happening at once which enables information to quickly propagate across the full school. As shown below, the result is a swift and decisive collective decision.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*JF4YlGQ9FiKgQ_gLPa1FwA.jpeg" /><figcaption>Fish Schools Make Rapid Smart Decisions</figcaption></figure><p>In this way, schooling fish can make<a href="https://www.pnas.org/doi/epdf/10.1073/pnas.1007102108"> rapid and effective decisions</a> across large populations, even when all members have limited information. Such a skill would be even more powerful for large human groups. After all, the problems faced by groups of people are significantly more complex and involve far more perspectives. This begs the question — could large human groups deliberate in real-time with the efficiency of fish schools and quickly reach optimized decisions?</p><p>For years this goal seemed impossible. That’s because <strong>conversational deliberations</strong> have been shown to be most productive<a href="https://www.jstor.org/stable/2786271"> in small groups of 4 to 7 people</a> and quickly degrade as groups grow larger. This is because the “airtime per person” gets progressively squeezed and the wait-time to respond to others steadily increases. By 12 to 15 people, the conversational dynamics change from thoughtful debate to a series of monologues that become increasingly disjointed. By 20 people, the dialog ceases to be a conversation at all. This problem seemed impenetrable until recent advances in Generative AI opened up new solutions.</p><p>The resulting technology is called <a href="https://arxiv.org/abs/2309.03220">Conversational Swarm Intelligence</a> and it promises to allow groups of almost any size (200, 2000, or even 2 million people) to discuss complex problems in real-time and quickly converge on solutions with significantly amplified intelligence. The first step is to divide the population into small subgroups, each sized for thoughtful dialog. For example, a 1000-person group could be divided into 200 subgroups of 5, each routed into their own chat room or video conferencing session. Of course, this does not create a single unified conversation — it creates 200 parallel conversations.</p><p>As described above, fish schools solve this problem by having local groups overlap, allowing information to quickly propagate across the full population. Unfortunately, we humans did not evolve with the ability to be in multiple conversations at once. In fact, if we try to pay attention to two conversations, we immediately get confused and can’t focus on either. This is commonly called the “<a href="https://en.wikipedia.org/wiki/Cocktail_party_effect">cocktail party problem</a>” because it happens often when small groups gather within earshot of each other. If you try to pay attention to a neighboring conversation you immediately lose track of the discussion you are in.</p><p><strong>So how can we overcome this human limitation?</strong></p><p>The technology of CSI solves this problem by inserting LLM-powered “<a href="https://arxiv.org/ftp/arxiv/papers/2309/2309.12366.pdf">Conversational Surrogate Agent</a>s” into each subgroup. These AI agents are tasked with distilling the real-time human insights within its group and sharing those insights with Surrogate Agents in one or more other groups. The receiving agents express the insights inside their own groups as natural first-person dialog. In this way, each subgroup is given an <em>artificial member</em> that participates seamlessly in the conversation while ensuring information freely propagates across the full population.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hlYu5FmizvAa1R33n3cYtw.gif" /></figure><p>A variety of recent studies suggest the approach is effective. For example, a<a href="https://arxiv.org/ftp/arxiv/papers/2309/2309.12366.pdf"> 2023 study</a> was conducted at<a href="https://www.thinkscape.ai/blog/carnegie-mellon-business-school-magazine-features-article-on-collective-superintelligence"> Carnegie Mellon</a> University to compare real-time deliberations among approximately 50 people in traditional chatrooms versus conversational swarms. When using the CSI structure, human groups with interconnected <a href="https://patents.google.com/patent/US11949638B1/en">conversational surrogate agents</a> were able to hold more coherent conversations that quickly converged on solutions. In addition, each individual was found to contribute 50% more content (on average) than participants using traditional methods.</p><p><strong>But does this amplify group intelligence?</strong></p><p>To explore this,<a href="https://unanimous.ai/wp-content/uploads/2024/04/Collective-Superintelligence-IQ-Paper-ICEIS-2024-official.pdf"> a 2024 study</a> by researchers at Carnegie Mellon and Unanimous AI tested the ability of networked groups to take IQ tests as a deliberative “hive mind.” Results showed that groups of 35 people, who averaged an IQ of 100 (the 50th percentile), could score an effective IQ of 128 (the 97th percentile) when using a CSI platform called<a href="https://thinkscape.ai/"> Thinkscape</a>. Although this study used conversational groups of only 35 participants, other recent studies have tested groups<a href="https://ieeexplore.ieee.org/document/10427807"> up to 250</a> with success.</p><figure><img alt="Thinkscape software was shown to amplify the IQ of networked human groups in a published academic study (Carnegie Mellon / Unanimous AI)" src="https://cdn-images-1.medium.com/max/619/1*BE1YEXiUsRdVusb4t_VNZw.jpeg" /><figcaption>Read the full academic study <a href="https://unanimous.ai/wp-content/uploads/2024/04/Collective-Superintelligence-IQ-Paper-ICEIS-2024-official.pdf">here</a></figcaption></figure><p>While the above studies used text conversations, the core methods of CSI can be deployed for voice or video meetings, enabling groups of potentially any size to hold coherent real-time conversations that facilitate problem solving and amplify group intelligence. The technology is also scalable, able to extend not just to hundreds of individuals <a href="https://patents.google.com/patent/US11949638B1/en">but thousands and potentially millions</a>, all in real-time. This has the potential to enhance a wide range of fields from enterprise collaboration and <a href="https://www.thinkscape.ai/blog/which-mixed-reality-headset-will-gain-mass-appeal-the-meta-quest-3-or-apple-vision-pro-we-asked-a-collective-intelligence">market research</a> to civic engagement and deliberative democracy.</p><p>In addition, recent research by my team at Unanimous AI has enabled large human groups to answer questions in through a unified personified avatar (which I call a <em>pluribus avatar</em>) that expresses the collective sentiment of the group in real-time. This can enable you, for example, to interview a thousand people in real-time, as if holding a conversation with collective.</p><p>Here is an example that was recently done using <a href="https://www.thinkscape.ai/">Thinkscape </a>with 113 randomly selected members of the public. The group was asked to hold a <a href="https://arxiv.org/abs/2309.12366">real-time conversational deliberation</a> about whether the U.S. should build new nuclear power plants. The first step was to ask the group the discuss their concerns regarding nuclear energy. After only six minutes of debate among the randomly selected participants, the AI-powered system produced this real-time representation of the group’s collective sentiment:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fplayer.vimeo.com%2Fvideo%2F979266860%3Fapp_id%3D122963&amp;dntp=1&amp;display_name=Vimeo&amp;url=https%3A%2F%2Fvimeo.com%2F979266860%3Fshare%3Dcopy&amp;image=https%3A%2F%2Fi.vimeocdn.com%2Fvideo%2F1894168753-0ac58aaea8e348ee40cb63b9e721919a622e8fe5c0eb2eed278df19783bb3a58-d_1280&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=vimeo" width="1912" height="1076" frameborder="0" scrolling="no"><a href="https://medium.com/media/d086ada1cb9e2e298ce2f6410f0bb86a/href">https://medium.com/media/d086ada1cb9e2e298ce2f6410f0bb86a/href</a></iframe><p>In the longer term, this approach could be used at extremely large scale to create a <a href="https://venturebeat.com/ai/the-promise-of-collective-superintelligence/">Collective Superintelligence </a>that can solve our most difficult problems while remaining inherently aligned with human values, morals, wisdom, and sensibilities. In theory, we could use CSI technology to enable millions of individuals to “think together” as a global <a href="https://www.brainyquote.com/quotes/louis_b_rosenberg_818625">brain-of-brains</a>. For me, this is preferable to allowing a purely digital superintelligence to make decisions for humanity, as I don’t believe that purely digital AI systems will maintain human interests or sensibilities over time.</p><p><a href="https://www.linkedin.com/in/louis-rosenberg/"><strong>Louis Rosenberg, PhD</strong></a> is a longtime researcher in the fields of artificial intelligence, collective intelligence, and mixed reality. He was formerly a professor at California Polytechnic and is currently the CEO and Chief Scientist of Unanimous AI. His latest book about AI, <a href="https://www.amazon.com/Our-Next-Reality-AI-powered-Metaverse/dp/1399812246">Our Next Reality</a> was recently released by Hachette (2024).</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=eade7f60c292" width="1" height="1" alt=""><hr><p><a href="https://medium.com/predict/how-to-build-a-collective-mind-that-speaks-for-humanity-in-real-time-eade7f60c292">How to build a Collective Mind that speaks for humanity in real-time.</a> was originally published in <a href="https://medium.com/predict">Predict</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The “AI Manipulation Problem” is urgent and not being addressed]]></title>
            <link>https://medium.com/predict/the-ai-manipulation-problem-is-urgent-and-not-being-addressed-ede0dd5e0b3e?source=rss-46e426f3cb75------2</link>
            <guid isPermaLink="false">https://medium.com/p/ede0dd5e0b3e</guid>
            <category><![CDATA[politics]]></category>
            <category><![CDATA[marketing]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Louis Rosenberg, PhD]]></dc:creator>
            <pubDate>Sun, 19 May 2024 20:05:37 GMT</pubDate>
            <atom:updated>2024-05-21T14:00:26.777Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*aS53lDuhBzhihFVmXy3k8A.jpeg" /><figcaption>The <a href="https://arxiv.org/abs/2306.11748">AI Manipulation Problem</a> is urgent and not being addressed</figcaption></figure><p>Our lives will soon be filled with<a href="https://venturebeat.com/ai/the-profound-danger-of-conversational-ai/"> conversational AI agents</a> designed to help us at every turn, anticipating our wants and needs so they can feed us tailored information and perform useful tasks on our behalf. They will do this using an extensive store of personal data about our individual interests and hobbies, backgrounds and aspirations, personality traits and political views — all with the goal of making our lives “more convenient.”</p><p>These agents will be extremely skilled. Just this week,<a href="https://venturebeat.com/ai/openai-announces-new-free-model-gpt-4o-and-chatgpt-for-desktop/"> Open AI released GPT-4o</a>, their next generation chatbot that can read human emotions. It can do this not just by reading sentiment in the text you write, but also by assessing the inflections in your voice (if you speak to it through a mic) and by using your facial cues (if you interact through video).</p><p><strong>This is the future of computing and it’s coming fast.</strong></p><p>Just this week, Google announced<a href="https://venturebeat.com/ai/google-takes-on-gpt-4o-with-project-astra-an-ai-agent-that-understands-dynamics-of-the-world"> Project Astra</a> — short for advanced seeing and talking responsive agent. The goal is to deploy an assistive AI that can interact conversationally with you while understanding what it sees and hears in your surroundings. This will enable it to provide interactive guidance and assistance in real-time.</p><p>And last week, Open AI’s Sam Altman<a href="https://www.technologyreview.com/2024/05/01/1091979/sam-altman-says-helpful-agents-are-poised-to-become-ais-killer-function/"> told MIT Technology Review</a> that the killer app for AI is assistive agents. In fact, he predicted everyone will want a personalized agent that acts as <em>“a super-competent colleague that knows absolutely everything about my whole life, every email, every conversation I’ve ever had,”</em> all captured so it can take useful actions on your behalf.</p><p><strong>What could possibly go wrong?</strong></p><p>As I wrote <a href="https://venturebeat.com/ai/why-generative-ai-is-more-dangerous-than-you-think/">in VentureBeat</a> last year, there is a significant risk that AI agents can be misused in ways that compromise human agency. In fact, I believe<a href="https://futureofmarketinginstitute.com/the-metaverse-from-marketing-to-mind-control/"> <em>targeted manipulation</em></a> is the single most dangerous threat posed by AI in the near future, especially when these agents become embedded in mobile devices. After all, mobile devices are the gateway to our digital lives, from the news and opinions we consume to every email, phone call, and text message we receive. These agents will monitor our information flow, learning intimate details about our lives, while also filtering the content that reaches our eyes.</p><p>Any system that monitors our lives and<a href="https://venturebeat.com/ai/snoop-dogg-sentient-ai-and-the-arrival-mind-paradox/"> mediates the information we receive</a> is a vehicle for interactive manipulation. To make this even more dangerous, these AI agents will use the cameras and microphones on our mobile devices to <em>see what we see</em> and <em>hear what we hear</em> in real-time. This capability (enabled by <em>Multimodal Large Language Models</em>) will make these agents extremely useful — able to react to the sights and sounds in your environment without you needing to ask for their guidance. This capability could also be used to trigger<a href="https://www.researchgate.net/publication/369355910_The_Manipulation_Problem_Conversational_AI_as_a_Threat_to_Epistemic_Agency"> targeted influence</a> that matches the precise activity or situation you are engaged in.</p><p>For many people, this level of tracking and intervention sounds creepy and yet, I predict they will<a href="https://venturebeat.com/ai/why-generative-ai-is-more-dangerous-than-you-think/"> embrace the technology</a>. After all, these agents will be designed to make our lives better, whispering in our ears as we go about our days, ensuring we don’t forget to pick up our laundry when walking down the street, tutoring us as we learn new skills, even coaching us in social situations to make us seem smarter, funnier, or more confident.</p><p>This will become an arms race among tech companies to<a href="https://venturebeat.com/ai/2024-will-be-the-year-of-augmented-mentality"> <em>augment our mental abilities</em></a> in the most powerful ways possible. And those who choose not to use these features will quickly feel disadvantaged. Eventually, it will not even feel like a choice. This is why I regularly predict that adoption will be extremely fast, becoming ubiquitous by 2030.</p><p><strong>So, why not embrace an <em>augmented mentality?</em></strong></p><p>As I wrote about in my new book,<a href="https://www.amazon.com/Our-Next-Reality-AI-powered-Metaverse/dp/1399812246"> Our Next Reality</a> — AI agents will give us mental superpowers, but we cannot forget these are products designed to make a profit. And by using these products, we will allow corporations to whisper in our ears (and flash images in our eyes) that guide us, coach us, educate us, caution us, and prod us throughout our days. In other words, we will allow AI agents to <em>influence our thoughts</em> and <em>guide our behaviors</em>. When used for good, this could be an amazing form of empowerment, but when abused it could easily become the<a href="https://www.researchgate.net/publication/368492998_The_Metaverse_and_Conversational_AI_as_a_Threat_Vector_for_Targeted_Influence"> ultimate tool of persuasion</a>.</p><p>This brings me to the “<a href="https://www.iiis.org/DOI2023/HA408FU/">AI Manipulation Problem</a>” — the fact that <em>targeted influence</em> delivered by conversational agents is potentially far more effective than traditional content. If you want to understand why, just ask any skilled salesperson. They know that the best way to coax someone into buying a product or service (even one they don’t need) is not to hand them a brochure, but to engage them in dialog. A good salesperson will start with friendly banter to “size you up” and lower your defenses. They will then ask questions to surface any reservations you may have. And finally, they will customize their pitch to overcome your concerns, using carefully chosen arguments that best play on your needs or insecurities.</p><p>The reason<a href="https://ieeexplore.ieee.org/document/10099167/"> AI manipulation</a> is such a significant risk is that AI agents will soon be able to pitch us interactively and they will be significantly more skilled than any human salesperson (see<a href="https://www.youtube.com/watch?v=IsE_Pas2OQU"> video example</a>). This is not only because these agents will be trained to use sales tactics, behavioral psychology, cognitive biases, and other tools of persuasion, but they will also be armed with far more information about us than any salesperson. In fact, if the agent is your “personal assistant,” it could know more about you than any human ever has. (For a depiction of AI assistants in the near future, read my 2021 short story <a href="https://www.metastellar.com/fiction/metaverse-2030/">Metaverse 2030</a> or see video below).</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FXoO8X0LAC8Q%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DXoO8X0LAC8Q&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FXoO8X0LAC8Q%2Fhqdefault.jpg&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/96b58d006c7e9b1b93df0ad31354f96d/href">https://medium.com/media/96b58d006c7e9b1b93df0ad31354f96d/href</a></iframe><p>In addition, these AI agents will be able to read your emotions with greater precision than any human — detecting “misexpressions” on your face and in your voice that no salesperson would ever notice. And if we don’t put regulation in place, they will document your reactions to every pitch they make, assessing what works best on YOU personally — offers for discounts, fear of missing out, logical arguments, emotional appeals — they will learn over time exactly <a href="https://www.barrons.com/articles/conversational-ai-will-learn-to-push-your-buttons-manipulation-problem-c9f797e8">how to push your buttons</a>.</p><p>Even the way AI agents appear will be optimized over time to ensure they maximize their impact on you personally. Their hair and eye color, clothing and speaking style, even age, gender, and race could all be chosen based on what is most likely to influence you best. I wrote about this a decade ago in my sci-fi <a href="https://www.amazon.com/gp/product/B008C8YM10/ref=dbs_a_def_rwt_bibl_vppi_i1">graphic novel UPGRADE</a> in which virtual spokespeople (called “Spokegens”) were deployed to target each individual. The AI agents would gradually adjust their appearance over time based on what works best on each person. The protagonist noticed that his agent became more and more sexualized until it always showed up looking like a prostitute — all so it could most efficiently sell him “food bars” and other products. I had intended this as irony back in 2012, but now it’s about to come true.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*G5A8JXpBPE6FNzpvS5eSIQ.jpeg" /><figcaption>Image from UPGRADE sci-fi graphic novel about artificial agents</figcaption></figure><p>From a technical perspective, the manipulative danger of AI agents can be summarized in two simple words — “feedback control”. That’s because a conversational agent can be given an “influence objective” and will then <a href="https://openaccess.cms-conferences.org/publications/book/978-1-958651-75-9/article/978-1-958651-75-9_13">work interactively to optimize the impact</a> of that influence on an specific person through direct conversation. It can do this by expressing a point, reading the reactions (both content and emotion)in your words, your vocal inflections, and your facial expressions, and then adapt its influence tactics (both its language and strategic approach) to overcome your objections and convince you of whatever it was asked to deploy.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/803/0*poz4QQoOjYJ0HdzX" /></figure><p>A<a href="https://www.iiis.org/CDs2023/CD2023Summer/papers/HA408FU.pdf"> control system for human manipulation</a> is shown above. From a conceptual perspective, it’s not very different than control systems used in <strong>heat seeking missiles</strong>. They detect the heat signature of an airplane and correct in real-time if they are not aimed in exactly the right direction, homing in until they hit their target. Unless regulated, conversational agents will be able to do the same thing, but the missile is<a href="https://venturebeat.com/ai/why-generative-ai-is-more-dangerous-than-you-think/"> a piece of influence</a>, and the <strong>target is you</strong>. And, if the influence is misinformation, disinformation, or propaganda the danger is extreme. For these reasons, regulators need to greatly limit targeted interactive influence.</p><p><strong>But are these technologies coming soon?</strong></p><p>I am confident that conversational agents<a href="https://medium.com/predict/the-age-of-augmented-mentality-d35b9fc9c24f"> will impact all our lives</a> within the next two to three years. After all, Meta, Google, and Apple have all made announcements that point in this direction. For example, Meta recently launched a new version of their<a href="https://about.fb.com/news/2024/04/new-ray-ban-meta-smart-glasses-styles-and-meta-ai-updates/"> Ray-Ban glasses</a> powered by AI that can process video from the onboard cameras, giving you guidance about items the AI can see in your surroundings. Apple is also pushing in this direction, announcing a<a href="https://venturebeat.com/ai/apple-researchers-achieve-breakthroughs-in-multimodal-ai-as-company-ramps-up-investments/"> Multimodal Large Language Model</a> that could give eyes and ears to Siri. As I wrote about here in<a href="https://venturebeat.com/ai/2024-will-be-the-year-of-augmented-mentality/"> VentureBeat</a>, I believe cameras will soon be included on most high-end earbuds to allow AI agents to always see what we’re looking at. As soon as these products are available to consumers, adoption will happen quickly. They will be useful.</p><p>Whether you are looking forward to it or not, the fact is — Big Tech is racing to put artificial agents into your ears (and soon our eyes) so they will guide us everywhere we go. There are very positive uses of these technologies that will make our lives better. At the same time, these same superpowers could easily be deployed as<a href="https://arxiv.org/abs/2306.11748"> agents of manipulation</a>.</p><p><strong>How do we address this risk?</strong></p><p>The need is clear — regulators <strong>must take rapid action</strong>, ensuring that the positive uses of AI agents are not hindered while protecting the public from abuse. I’ve been writing about the <a href="https://futurism.com/artificial-intelligence-isnt-intelligence-manipulating-humanity">AI Manipulation Problem since 2016</a> and have given multiple presentations to the FTC over the last few years, and thus far regulators have not put any protections in place.</p><p>And it’s not just the US that is ignoring this emerging problem — regulators around the world are still focused on “old school” risks of AI that involve their ability to generate traditional misinformation at scale. Those are real dangers, but the far greater risk is <a href="https://www.researchgate.net/publication/368492998_The_Metaverse_and_Conversational_AI_as_a_Threat_Vector_for_Targeted_Influence">targeted interactive manipulation</a>. And with the announcements last week from Open AI and Google, it’s now clear that our time is up — regulators NEED to act now. The short video below entitled <a href="https://www.privacylost.org/">Privacy Lost</a> (3 min) is designed to show the manipulation risk.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FIsE_Pas2OQU%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DIsE_Pas2OQU&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FIsE_Pas2OQU%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/95e67ff265ca6fea45791c6e899cdb16/href">https://medium.com/media/95e67ff265ca6fea45791c6e899cdb16/href</a></iframe><p>To address this, the first big step would be an <strong>outright ban</strong> (or very strict limits) on<a href="https://www.researchgate.net/publication/369355910_The_Manipulation_Problem_Conversational_AI_as_a_Threat_to_Epistemic_Agency"> interactive conversational advertising</a>. Interactive advertising is the “gateway drug” to conversational propaganda and misinformation. And if we don’t ban this tactic, it will quickly become an arms race, with Big Tech companies competing with each other to deploy the most “efficient” and “effective” conversational ads that drive users to make purchases. And the very same technologies that they optimize for driving products and services will be used to drive misinformation and propaganda, optimizing the ability to make you believe things that are not in your best interest.</p><p>The time for policymakers to address this is now.</p><p><a href="https://sites.google.com/view/louisrosenberg/bio"><strong>Louis Rosenberg, PhD</strong></a> is a American technologist in the fields of AI and XR. His new book,<a href="https://www.amazon.com/Our-Next-Reality-AI-powered-Metaverse/dp/1399812246"> Our Next Reality</a>, about the impact of AI on society was recently released by Hachette. He earned his PhD from Stanford, was a professor at California Polytechnic and is currently CEO of Unanimous AI.</p><p>This piece originally appeared in <a href="https://venturebeat.com/ai/agents-of-manipulation-the-real-ai-risk/">Venture Beat (5/17/24)</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ede0dd5e0b3e" width="1" height="1" alt=""><hr><p><a href="https://medium.com/predict/the-ai-manipulation-problem-is-urgent-and-not-being-addressed-ede0dd5e0b3e">The “AI Manipulation Problem” is urgent and not being addressed</a> was originally published in <a href="https://medium.com/predict">Predict</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Mixed Reality or Spatial Computing]]></title>
            <link>https://medium.com/predict/mixed-reality-or-spatial-computing-346e62148026?source=rss-46e426f3cb75------2</link>
            <guid isPermaLink="false">https://medium.com/p/346e62148026</guid>
            <category><![CDATA[metaverse]]></category>
            <category><![CDATA[spatial-computing]]></category>
            <category><![CDATA[mixed-reality]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[marketing]]></category>
            <dc:creator><![CDATA[Louis Rosenberg, PhD]]></dc:creator>
            <pubDate>Fri, 16 Feb 2024 23:09:10 GMT</pubDate>
            <atom:updated>2024-11-22T14:48:36.612Z</atom:updated>
            <content:encoded><![CDATA[<p><strong>The Language of Immersive Media — Past and Present</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/475/1*lH3sHZrL7lBzFaalXtedzQ.jpeg" /><figcaption>Virtual Reality as depicted by NASA (<a href="https://commons.wikimedia.org/wiki/File:THE_VIEW_(Virtual_Reality).jpg">circa 1992</a>)</figcaption></figure><p>Journalists keep asking me the same question these days: <strong><em>What’s the difference between Spatial Computing and Mixed Reality? </em></strong>As soon as I answer, it triggers even more questions about the language of immersive media and the various phrases in use today. To reduce confusion, I believe it’s helpful to review the history and <em>perceptual meaning</em> of core language.</p><p>Personally, I like Apple’s use of the phrase <strong><em>spatial computing</em></strong> and agree the Vision Pro is a <strong><em>spatial computer. </em></strong>Still, this has added to the confusion. After all, just about everyone calls the Vision Pro a 𝘮𝘪𝘹𝘦𝘥 𝘳𝘦𝘢𝘭𝘪𝘵𝘺 𝘩𝘦𝘢𝘥𝘴𝘦𝘵 except for Apple. Also, <a href="https://developer.apple.com/visionos/submit/">developers</a> are not allowed to describe Vision Pro apps as enabling <strong><em>virtual reality</em></strong>, <strong><em>mixed reality</em></strong>, or <strong><em>augmented reality</em></strong>. I appreciate the marketing reasons for this, and suspect these language restrictions will fade over time — but it seems like overkill for Apple to ban certain words.</p><p>Maybe I’m just nostalgic, but when I first started working in the field, the phrase <strong>virtual reality</strong> was only a few years old and was already generating major excitement. I was a young researcher conducting <a href="https://apps.dtic.mil/sti/citations/ADA289764">VR experiments</a> at NASA and the photo above was a poster in the lab where I was working. To me, it was a deeply inspiring image, capturing the full promise of the field.</p><p>Even more important, the human experience depicted in that NASA photo has been called <strong>virtual reality</strong> for almost 40 years. If you are a developer for the Vision Pro and you create a fully simulated immersive experience, is it really a sin to describe it as <strong>virtual reality</strong>? After all, the VR headset shown above is now in the Smithsonian. This is our history and culture.</p><p>Of course, the Vision Pro is far more sophisticated than the NASA headset above, not just because of its amazing fidelity but because it adds entirely new capabilities. The most significant capability is the power of the Vision Pro to seamlessly combine the real world with spatially projected virtual content to create a single unified experience — a <em>single perceptual reality</em>. This is called <a href="https://venturebeat.com/datadecisionmakers/augmented-reality-superhuman-abilities-and-the-future-of-medicine/">augmented reality</a> or <a href="https://venturebeat.com/virtual/2023-could-be-the-year-of-mixed-reality/">mixed reality</a> depending on the capabilities and both phrases have a long history in academia and industry.</p><p><strong>So, what’s the difference between AR and MR?</strong></p><p>This is probably <strong>the most misunderstood issue</strong> in the world of immersive media, so it’s worth taking a trip back in time to explain how today’s divide came to be. In the early days, only one phrase was needed, <em>Augmented Reality</em>, but its definition was diluted in the 2010’s as marketeers pushed simpler and simpler systems under that banner. I suspect the pendulum will swing back in the future, but for now both phrases are helpful. To appreciate why, we need to dig into the <strong>perceptual requirements</strong> for augmenting a user’s reality in a convincing and authentic way.</p><p>As background, I began working on merging the real and the virtual in 1991 before we had language to describe such a combined experience. My early research at Stanford, NASA and the US Air Force aimed to explore the basic psychophysical requirements needed to create a <em>unified perceptual reality</em> of the physical and virtual. I called this new pursuit “<a href="https://dl.acm.org/doi/10.5555/221788">design for perception</a>” (not very catchy) and called the virtual objects “<a href="https://www.spiedigitallibrary.org/conference-proceedings-of-spie/2057/1/Virtual-fixtures-as-tools-to-enhance-operator-performance-in-telepresence/10.1117/12.164901.short">perceptual overlays</a>” (also not catchy). This is how I described the core perceptual requirements for creating an authentic and believable mixed reality back then:</p><p><strong>Perceptual Overlays</strong>: <em>“I do not want to impose any restrictions on the nature or content of a perceptual overlay, other than to require that a natural and predictable relation exists between an operator’s neuromotor activities (efference) and the subsequent changes in the sensations included in the perceptual overlay (afference). This single restriction upon perceptual overlays is sufficient to encourage distal attribution to occur. As a result of distal attribution, the operator will accept the overlaid percepts as real and tangible parts of the ambient reality. In other words, the operator will achieve a sense of presence with respect to the overlaid perceptual information. Thus, the virtual percepts will be accepted as genuine features or properties of the environment in which the operator is working.”</em> (<a href="https://search.worldcat.org/title/123253939">Stanford, Rosenberg, 1993</a>)</p><p>That excerpt is quite a mouthful, but the key phrase is <em>“distal attribution.”</em> It refers to the process in which your brain receives a new piece of perceptual information and integrates it into your reality as an authentic part of your surroundings. Thus, the guidance above states that to have a virtual object be perceived as a genuine part of your <em>ambient reality</em>, we need to enable <a href="https://link.springer.com/chapter/10.1007/978-3-030-89906-6_1">interactive experiences</a>. Passive viewing is not sufficient because it does not close the mental loop between afference and efference (i.e., action and reaction). This is why 360 degree VR movies filmed from a fixed location don’t quite seem like reality, even if you have depth perception and can turn your head. Without the ability to interact by moving or reaching, your brain categorizes it as a façade over your field of view, not your true reality.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TdDDgo1Al-NhD2ByvW84lQ.jpeg" /><figcaption><strong>Psychophysical Model for merging the Real and Virtual </strong>(<a href="https://search.worldcat.org/title/123253939">Rosenberg 1994</a>)</figcaption></figure><p>Identifying <strong>distal attribution </strong>as a core requirement for creating a unified mixed reality was a helpful insight, but I soon confronted additional perceptual requirements that had to be met to merge the real and virtual in a convincing way. These requirements can be summarized as (a) <strong>accurate spatial alignment</strong> in full 3D, (b) <strong>simultaneous interactivity</strong> with both the real and the virtual, and (c) <strong>bidirectional interactivity</strong> between the real and the virtual. Let me address each:</p><p>First and foremost, the real and the virtual need to be <strong>spatially aligned in 3D space </strong>with sufficient precision that the flaws are beyond the limits of human perception (called “Just Noticeable Difference” or JND in the field of psychophysics). Even subtle flaws destroy the illusion and your brain will perceive the real and virtual as separate percepts, not one reality. The two realms also need to be aligned in time, which sounds simple but with computer lag (especially in the old days) is often just as difficult.</p><p>Next, both realms need to be <strong>simultaneously interactive. </strong>This means, users need to be able to engage naturally with the real and the virtual at the same time, cementing the illusion that the digital content is an authentic part of the physical surroundings. It turns out that <a href="https://apps.dtic.mil/sti/tr/pdf/ADA296190.pdf">physical engagement</a> (especially manual interaction) may be the most convincing of all. For example, if you reach out and grab a real object and move it naturally with respect to a virtual object, the illusion is firmly cemented. I suspect this is because our eyes are easy to fool, so our brains are more skeptical, but when you can manipulate both worlds manually and maintain perceptual consistency, your brain snaps the two worlds together as one.</p><p>And finally, the real and virtual need to engage each other (<strong>bi-directionally interactive</strong>), because without that consistency the illusion is compromised. If you grab a virtual book and place it on a real table and it passes through, it’s <a href="https://apps.dtic.mil/sti/citations/ADA296190">not perceived as a single reality</a> and suspension of disbelief is lost. But if you maintain the illusion, your brain buys in and you find yourself in a single reality, a mixed reality that is deeply convincing — so convincing, you stop thinking about what’s real and what’s virtual. It’s just one reality.</p><p>The <a href="https://spectrum.ieee.org/history-of-augmented-reality">Virtual Fixtures platform</a> developed at Wright Patterson Air Force Base (1991–1994) met these perceptual requirements for the first time, enabling users to experience a unified reality where virtual objects (called <em>virtual fixtures</em>) were perceived as authentic additions to the physical world. It was very crude by today’s standards, but because it enabled distal attribution, manual interactivity, bi-directional interactivity, and real-time feedback with 3D haptics and 3D audio, it created the first mixed reality experiences for users (i.e., a <em>unified perceptual reality</em>). The system is shown below during an experiment in 1993 in which spatially registered perceptual overlays were employed to <a href="https://www.spiedigitallibrary.org/conference-proceedings-of-spie/2057/1/Virtual-fixtures-as-tools-to-enhance-operator-performance-in-telepresence/10.1117/12.164901.short">help users perform</a> dexterous real-world tasks:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/277/1*MurjNHOuRJQ6pnkIYD3_9w.gif" /><figcaption><strong>Virtual Fixtures Platform — mixed reality research</strong> (USAF 1993, Rosenberg)</figcaption></figure><p>I bring up these perceptual requirements <em>(distal attribution, 3D registration, simultaneous interactivity, and bi-directional interactivity)</em> because they help explain why the Vision Pro is such a remarkable device. It’s the first product to achieve these capabilities with extreme precision, fostering a unified perceptual reality and maintaining the illusion for extended periods without flaws taking users out of the experience. The Air Force prototype shown above put many constraints on the user and could only maintain the illusion for specific tasks.</p><p>Now, back to language — how did we end up with two phrases <em>augmented reality</em> and <em>mixed reality</em> that refer to merging the real and virtual? When I shot the video above, no language yet existed to refer to this field of research. Fortunately, the phrase augmented reality soon emerged from Boeing which clearly reflected the goals — to add virtual content to a real environment that is so accurately aligned and naturally integrated, the <a href="https://link.springer.com/chapter/10.1007/978-3-030-89906-6_1">two worlds merge together in your mind</a>. And for two decades, that’s what AR meant (while simpler devices that merely embellished or annotated your field of view without distal attribution were called <a href="https://en.wikipedia.org/wiki/Head-up_display">head-up displays</a>).</p><p>Then in<em> 2013 </em><a href="https://en.wikipedia.org/wiki/Google_Glass"><em>Google Glass</em></a><em> </em>happened<em>. </em>I deeply respect the product and believe it was ahead of its time. Unfortunately, the media incorrectly referred to it as augmented reality. It was not. It didn’t enable virtual content to be placed into the real world in a way that was immersive, spatially registered, or interactive. Instead, it was what we now call<strong> smart glasses</strong>, which are profoundly useful and will become <a href="https://venturebeat.com/ai/2024-will-be-the-year-of-augmented-mentality/">even more useful as AI gets integrated into these products</a>, but it wasn’t AR.</p><p>Still, the phrase <a href="https://link.springer.com/chapter/10.1007/978-3-030-89906-6_1">augmented reality</a> got watered down during the 2010’s, not only because of Google Glass but because the makers of smartphones were pushing simple visual overlays as “augmented reality,” even though they were not immersive and lacked genuine <strong>3D registration </strong>with the real world. They also lacked <strong>user interactivity</strong> and <strong>bi-directional interactivity</strong> between the real and virtual. This was before LiDAR and other spatial mapping technologies were added to phones, enabling increasing levels of 3D spatial registration and interactivity. Today’s phones are much better, but the phrase AR got diluted.</p><p>As a result, when Microsoft launched the first commercial product (the HoloLens) that fostered a unified perceptual reality, they likely felt it was necessary to distance the device from the language of prior systems. I suspect this is why Microsoft, upon launching HoloLens, focused their marketing on the words <em>mixed reality</em>. The phrase had been around since the 1990’s, but it was with the HoloLens launch that the language really took off. It basically came to mean creating a <em>genuine</em> augmented reality, or as I would put it more rigorously <em>— merging the real and the virtual into a unified perceptual reality.</em></p><p>And so, we now have <strong>two terms</strong> that describe different levels of augmenting a user’s surroundings with spatially registered virtual content. To clarify the difference between AR, MR, and VR, we can look at <a href="https://www.gao.gov/assets/gao-22-105541.pdf">definitions that were published in 2022</a> by the U.S. Government Accountability Office (GAO). I assume the GAO cares about these differences to clarify if government contracts are paying for VR, AR, or MR devices. To address this, the GAO put out a public document that featured the simple image below. I like their framing because it gets at the key issue of interactivity which is so important to distal attribution:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/488/1*gBz1IdVrQyM1-6BKCuIL-g.jpeg" /><figcaption><strong>VR vs MR vs AR as defined by U.S. GAO</strong> (<a href="https://www.gao.gov/assets/gao-22-105541.pdf">GAO-22–105541</a>)</figcaption></figure><p>It’s worth noting that the difference between AR and MR has nothing to do with the hardware and everything to do with the perceptual experience. I bring this up because many people incorrectly believe that AR hardware refers to glasses with transparent screens you can peer through, and MR hardware refers to headsets that use “passthrough cameras” to capture the real world and display it to the user on internal screens. This is not true. For example, passthrough cameras were used back in 1992 when developing the Virtual Fixtures platform. This design choice was made because it made it easier to <a href="https://spectrum.ieee.org/history-of-augmented-reality">register the real and virtual</a> with higher precision, not because it changed the user experience. And besides, simple phone-based AR also uses cameras, so that is not the differentiator.</p><p>This leads me back to the Apple Vision Pro — it is a mixed reality headset, not because it uses passthrough cameras, but because it enables users to experience a unified perceptual reality of the real and virtual. And because <a href="https://www.youtube.com/watch?v=El16ZDfd6Yo">mixed reality</a> is the superset capability, the Vision Pro can also provide simpler augmented reality experiences and fully simulated virtual reality experiences. And for all three (VR, AR, and MR), the Vision Pro can amaze consumers with immersive experiences that far exceed any device built at any price. It’s a true innovation for Apple and an achievement for the team of engineers who made it happen.</p><figure><img alt="Photo of Louis Rosenberg developing the mixed reality at Air Force Research Laboratory (1992) along side Rosenberg using mixed reality in 2022." src="https://cdn-images-1.medium.com/max/1024/1*i5MQoAsk0SzCwx56LCi03Q.jpeg" /><figcaption><strong>Left:</strong> Rosenberg researching mixed reality at Air Force Research Laboratory in 1992. <strong>Right:</strong> Rosenberg interacting with mixed reality thirty years later with the Meta Quest Pro in 2022.</figcaption></figure><p>The Vision Pro also enables new abilities that are entirely unique, including a <strong>spatial operating system</strong> (<a href="https://en.wikipedia.org/wiki/VisionOS">visionOS</a>) that breaks new ground by using a user’s gaze direction for input. In other words, I agree that the Vision Pro is not only a mixed reality headset, but also a <em>spatial computer</em> and frankly, a work of art. I also believe that <a href="https://venturebeat.com/ai/the-language-wars-of-2024-apple-vs-reality/"><em>spatial computing</em></a> is a great overarching term for AR, MR, and VR, along with other immersive experiences such as 3D movies and telepresence. My only recommendation is that all companies embrace the historic and accepted language of the field. Spatial computing is a useful term, but so is augmented reality, mixed reality, and virtual reality, all three of which are part of our history and culture.</p><p><a href="https://sites.google.com/view/louisrosenberg/bio"><strong><em>Louis Rosenberg, PhD</em></strong></a><em> is a longtime researcher in the fields of virtual reality, augmented reality, and artificial intelligence. He is known for founding Immersion Corporation (IMMR: Nasdaq) in 1993 and Unanimous AI in 2014, and for developing mixed reality at Air Force Research Laboratory. His new book, </em><a href="https://www.amazon.com/Our-Next-Reality-AI-powered-Metaverse-ebook/dp/B0CL9BWXCJ"><em>Our Next Reality</em></a><em>, is available for preorder from Hachette.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=346e62148026" width="1" height="1" alt=""><hr><p><a href="https://medium.com/predict/mixed-reality-or-spatial-computing-346e62148026">Mixed Reality or Spatial Computing</a> was originally published in <a href="https://medium.com/predict">Predict</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Age of Augmented Mentality]]></title>
            <link>https://medium.com/predict/the-age-of-augmented-mentality-d35b9fc9c24f?source=rss-46e426f3cb75------2</link>
            <guid isPermaLink="false">https://medium.com/p/d35b9fc9c24f</guid>
            <category><![CDATA[augmented-reality]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[marketing]]></category>
            <dc:creator><![CDATA[Louis Rosenberg, PhD]]></dc:creator>
            <pubDate>Sun, 14 Jan 2024 20:57:47 GMT</pubDate>
            <atom:updated>2024-01-23T12:45:52.138Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*F4nxjwKZ9HFM4Ld-0Xt5UA.png" /><figcaption>Welcome to the Age of Augmented Mentality</figcaption></figure><p>In the near future, an AI assistant will make itself at home inside your ears, <a href="https://venturebeat.com/ai/the-profound-danger-of-conversational-ai/">whispering guidance</a> as you go about your daily routine. It will be an active participant in all aspects of your life, providing useful information as you browse the isles in crowded stores, take your kids to see the pediatrician, even when you grab a quick snack from a cupboard in the privacy of your own home. It will <a href="https://venturebeat.com/ai/2024-will-be-the-year-of-augmented-mentality/">mediate all of your experiences</a>, including your social interactions with friends, relatives, coworkers, and strangers.</p><p>Of course, the word “mediate” is a euphemism for allowing an AI to influence what you do, say, think and feel. Many people will find this notion creepy, and yet as a society we will accept this technology into our lives, allowing ourselves to be continuously <a href="https://www.researchgate.net/publication/369355910_The_Manipulation_Problem_Conversational_AI_as_a_Threat_to_Epistemic_Agency">coached by friendly voices</a> that inform us and guide us with such skill, we will soon wonder how we ever lived without the real-time assistance.</p><p>When I use the phrase “AI assistant,” most people think of old-school tools like Siri or Alexa that allow you to make simple requests through verbal commands. This is not the right mental model. That’s because <a href="https://ieeexplore.ieee.org/document/10099167">next-generation assistants</a> will include a new ingredient that changes everything — <em>context awareness.</em> This additional capability will allow these systems to respond not just to what you say, but to the sights and sounds that you are currently experiencing all around you, captured by cameras and microphones on AI-powered devices that you will wear on your body.</p><p>Whether you’re looking forward to it or not, context-aware AI assistants will hit society in 2024, and they will significantly change our world within just a few years, unleashing a flood of powerful capabilities along with a torrent of new risks to personal privacy and human agency. On the positive side, these assistants will provide valuable information everywhere you go, precisely coordinated with whatever you’re doing, saying, or looking at. The guidance will be delivered so smoothly and naturally, it will feel <a href="https://bigthink.com/the-future/ar-will-make-us-superhumans/">like a superpower</a> — a voice in your head that knows everything, from the specifications of products in a store window, to the names of plants you pass on a hike, to the best dish you can make with the scattered ingredients in your refrigerator. On the negative side, this ever-present voice could be highly persuasive, even manipulative as it assists you through your daily activities, especially if corporations use these trusted assistants to deploy targeted <a href="https://arxiv.org/abs/2306.11748">conversational advertising</a>.</p><p>The <a href="https://www.iiis.org/DOI2023/HA408FU/">risk of AI manipulation</a> can be mitigated, but it requires policymakers to focus on this critical issue, which thus far has been largely ignored. Of course, regulators have not had much time — the technology that makes context-aware assistants viable for mainstream use has only been available for less than a year. The tech is called <a href="https://arxiv.org/abs/2306.13549">Multi-Modal Large Language Models</a> and it is new class of LLMs that can accept as input not just text prompts, but also images, audio, and video. This is a major advancement, for multi-modal models have suddenly given AI systems their own eyes and ears and they will use these <em>sensory organs </em>to assess the world around us as they give guidance in real-time.</p><p>The first mainstream multi-modal model was ChatGPT 4 released by Open AI in March of 2023. The most recent major entry into this space was Google’s Gemini LLM announced just a few weeks ago. The most interesting entry (to me personally) is the multi-modal LLM from Meta called <a href="https://arxiv.org/abs/2309.16058">AnyMAL</a> that also takes in motion cues. This model goes beyond adding eyes and ears to the AI, and essentially adds a vestibular sense of movement. This could be used to create an AI assistant that doesn’t just see and hear everything you experience, it even considers your physical state of motion.</p><p>With this AI technology now available for consumer use, companies are rushing to build them into systems that can see and hear your life and guide you through your daily interactions. This means putting a camera, microphone, and motion sensors on your body in a way that can feed the AI model and allow it to provide context-aware assistance throughout your life. The most natural place to put these sensors is in glasses because that ensures the cameras are looking in the direction of your gaze. Stereo microphones on eyewear (or earbuds) can also capture the soundscape with spatial fidelity, allowing the AI to know the direction that sounds are coming from — like barking dogs, honking cars, and crying kids.</p><p>In my opinion, the company that is currently leading this space is Meta. Two months ago they began selling a new version of their <a href="https://about.fb.com/news/2023/09/new-ray-ban-meta-smart-glasses/">Ray-Ban smart glasses</a> that was configured to support advanced AI models. The big question I’ve been tracking is <em>when</em> they would roll out the software needed to provide context-aware AI assistance. That’s no longer an unknown — on December 12 they began providing early access to the AI features which include remarkable capabilities. In the release video, Mark Zuckerberg demonstrated asking the AI assistant to suggest a pair of pants that would match a shirt he was looking at. It replied with skilled suggestions. Similar guidance could be provided while cooking, shopping, traveling, and of course, socializing. And the assistance will be context aware, for example <a href="https://patents.google.com/patent/US7577522B2/en">reminding you to buy dog food</a> when you walk past a pet store.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*RDnMjPh_XGtFE5tw_JQRsQ.jpeg" /><figcaption>Meta Smart Glasses (Wikimedia Commons)</figcaption></figure><p>Another high-profile company is Humane which developed <a href="https://www.nytimes.com/2023/11/09/technology/silicon-valleys-big-bold-sci-fi-bet-on-the-device-that-comes-after-the-smartphone.html">a wearable pin</a> with cameras and microphones and also promises to provide AI-powered assistance to users. Their device starts shipping in early 2024 and likely will capture the imagination of hardcore tech enthusiasts. That said, I personally believe that glasses-worn sensors are more effective than body-worn sensors because they detect the direction you’re looking and because it can also include visual elements added to your line of sight. These visual elements are simple overlays today but over the next five years will become rich and immersive <a href="https://www.researchgate.net/publication/376809737_Mixed_Reality_the_Future_of_Computing">mixed reality</a> experiences.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/600/1*WPI4nJNpNc-JEe0Ku47CnA.png" /><figcaption>Humane Pin (Wikimedia Commons)</figcaption></figure><p>Regardless of whether these context-aware AI assistants are enabled by sensored glasses, earbuds, or pins, they will become widely adopted in the next few years. That’s because they will offer powerful features from real-time translation of foreign languages to historical content about the world around us. These tools may even enable a <a href="https://patents.google.com/patent/US7732694B2">fun mixed reality technology</a> that I invented as an AR researcher 18 years ago <em>— </em><a href="https://patents.google.com/patent/US7542816B2"><em>context aware media players</em></a> that adapt musical experiences to your surroundings, giving your life the look and feel of its own movie soundtrack. But most significant, these devices will provide real-time assistance during social interactions, reminding us of the names of coworkers we meet on the street, suggesting funny things to say during lulls in conversations, even warning us when the person we’re talking to is getting annoyed or bored based on subtle facial or vocal cues (even <a href="https://www.sciencedirect.com/science/article/pii/S1877042815055779">micro-expressions</a> that are not perceptible to humans but easily detectable by AI).</p><p>Yes, whispering AI assistants will make everyone seem more charming, more intelligent, more socially aware, and potentially more persuasive as they coach us in real time. And it will become an arms race, with assistants working to give us an edge while protecting us from the persuasion of others. As a lifetime researcher into the impacts of <a href="https://bigthink.com/the-future/metaverse-augmented-reality-danger/">AI and mixed reality</a>, I’ve been worried about this danger for decades. To raise awareness, a few years ago I published a short story entitled <a href="https://www.amazon.com/Spring-Into-SciFi-Jason-Lairamore-ebook/dp/B08XLSGMP6">Carbon Dating</a> about a fictional AI that whispers advice in people’s ears. In the story, an elderly couple has their first date, neither saying anything that’s not coached by AI. It might as well be the courting ritual of two digital assistants, not two humans, and yet this ironic scenario may soon become commonplace. To help the public and policymakers appreciate the risks, Carbon Dating was recently turned into a fun <a href="https://www.youtube.com/watch?v=XoO8X0LAC8Q">audio play called Metaverse 2030</a> by the UK’s Office of Data Protection Authority (ODPA).</p><p>Of course, the biggest risks are not AI assistants butting in when we chat with friends, family, and romantic interests. The biggest risks are how corporate or government entities could inject their own agenda, enabling powerful forms of <a href="https://www.mi-3.com.au/03-11-2023/presidential-executive-order-ai-risks-missing-biggest-emerging-threats-and-most-powerful">conversational influence</a> that target us with customized content generated by AI to <a href="https://futureofmarketinginstitute.com/why-generative-ai-is-a-new-form-of-media/">maximize its impact on each individual</a>. To educate the public about the manipulative risks of these personalized AI systems, a short film entitled <a href="https://www.privacylost.org/">Privacy Lost</a> was recently created by the Responsible Metaverse Alliance with funding from Minderoo Pictures. As shown below, the three-minute video make the potential for targeted influence very clear:</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FIsE_Pas2OQU%3Ffeature%3Doembed&amp;display_name=YouTube&amp;url=https%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DIsE_Pas2OQU&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FIsE_Pas2OQU%2Fhqdefault.jpg&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/95e67ff265ca6fea45791c6e899cdb16/href">https://medium.com/media/95e67ff265ca6fea45791c6e899cdb16/href</a></iframe><p><strong>Do we have a choice?</strong></p><p>For many people, the idea of allowing AI assistants to whisper in their ears is a <a href="https://bigthink.com/the-present/danger-conversational-ai/">creepy scenario</a> they intend to avoid. The problem is, once a significant percentage of users are being coached by powerful AI tools, those of us who reject the features will be at a disadvantage. In fact, AI coaching will likely become part of the basic social norms of society, with everyone you meet expecting that you’re being fed information about them in real-time as you hold a conversation. It could become rude to ask someone what they do for a living or where they grew up, because that information will simply appear in your glasses or be whispered in your ears. And when you say something clever or insightful, nobody will know if you came up with it yourself or if you’re just parroting the AI assistant in your head. The fact is, we are headed towards a new social order in which we’re not just influenced by AI, but effectively <em>augmented</em> in our mental and social capabilities by AI tools provided by corporations.</p><p>I call this technology trend “<strong>augmented mentality</strong>” and while I believe it is inevitable, I thought we had more time before AI products would be launched that guide our daily thoughts and behaviors. But with recent advancements like context-aware LLMs, there are no longer technical barriers. This is coming and it will likely lead to an arms race in which the titans of Big Tech battle for bragging rights on who can pump the <a href="https://techcrunch.com/2022/01/12/the-metaverse-will-be-filled-with-elves/">most powerful AI guidance</a> into your eyes and ears. And of course, this corporate push could create a dangerous digital divide between those who can afford these intelligence enhancing tools and those who cannot. Or worse, those who can’t afford a subscription fee could be pressured to accept sponsored ads delivered through aggressive AI-powered <a href="https://bigthink.com/the-future/playing-god-metaverse-mind-control-free-will/">conversational influence</a>.</p><p><strong>Is this really the future we want to unleash?</strong></p><p>We are about to live in a world where corporations can literally put voices in our heads that influence our actions and opinions. This is the <a href="https://cepa.org/article/worried-about-ai-consider-computers-that-converse/">AI manipulation problem</a> and it is so worrisome, we urgently need <a href="https://www.researchgate.net/publication/369355910_The_Manipulation_Problem_Conversational_AI_as_a_Threat_to_Epistemic_Agency">aggressive regulation</a> of AI systems that “close the loop” around individual users in real-time, sensing our personal actions while imparting custom influence. Unfortunately, the recent <a href="https://www.whitehouse.gov/briefing-room/presidential-actions/2023/10/30/executive-order-on-the-safe-secure-and-trustworthy-development-and-use-of-artificial-intelligence/">Executive Order on AI</a> from the White House did not address this issue, while the <a href="https://www.europarl.europa.eu/news/en/headlines/society/20230601STO93804/eu-ai-act-first-regulation-on-artificial-intelligence">EU’s recent AI ACT</a> only touched on it tangentially. And yet, consumer products are about to flood the market designed to guide us throughout our lives.</p><p>As we dive into 2024, I sincerely hope that policymakers around the world shift their focus to the dangers of AI-powered <a href="https://www.iiis.org/DOI2023/HA408FU/">conversational influence</a>, especially when delivered by context-aware assistants. If they address these issues thoughtfully, consumers can have the benefits of AI guidance without it driving society down a dangerous path. For details on the potential regulatory solutions against AI manipulation, I recommend my new book <a href="https://www.amazon.com/Our-Next-Reality-Preparing-AI-powered/dp/1399812246/ref=sr_1_1">Our Next Reality</a> that just went live for preorder on Amazon.</p><p>This piece originally appeared <a href="https://venturebeat.com/ai/2024-will-be-the-year-of-augmented-mentality/">in VentureBeat </a>(1/1/2024)</p><p><a href="https://sites.google.com/view/louisrosenberg/bio"><strong><em>Louis Rosenberg, PhD</em></strong></a><em> has been a pioneering researcher in the fields of VR, AR, and AI for over thirty years. He is known for founding Immersion Corporation (IMMR: Nasdaq) and </em><a href="https://unanimous.ai/"><em>Unanimous AI</em></a><em>, and for developing the first </em><a href="https://spectrum.ieee.org/history-of-augmented-reality"><em>mixed reality</em></a><em> system at Air Force Research Laboratory. His new book, </em><a href="https://unanimous.ai/wp-content/uploads/2024/01/Our-Next-Reality-Press-Release-1.pdf"><em>Our Next Reality</em></a><em>, comes out in March from Hachette.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d35b9fc9c24f" width="1" height="1" alt=""><hr><p><a href="https://medium.com/predict/the-age-of-augmented-mentality-d35b9fc9c24f">The Age of Augmented Mentality</a> was originally published in <a href="https://medium.com/predict">Predict</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Artificial Superintelligence and the Arrival Mind Paradox]]></title>
            <link>https://medium.com/predict/artificial-superintelligence-and-the-arrival-mind-paradox-3958590f50c5?source=rss-46e426f3cb75------2</link>
            <guid isPermaLink="false">https://medium.com/p/3958590f50c5</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[social-media]]></category>
            <category><![CDATA[chatgpt]]></category>
            <category><![CDATA[marketing]]></category>
            <dc:creator><![CDATA[Louis Rosenberg, PhD]]></dc:creator>
            <pubDate>Mon, 30 Oct 2023 14:02:47 GMT</pubDate>
            <atom:updated>2023-10-30T14:02:47.486Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*DcKOHUEaHI5xjhTFxilOuQ.png" /><figcaption>Superintelligence, Sentient AI, and the Arrival Mind Paradox</figcaption></figure><p>There’s endless conversation these days about the “<strong>existential risks</strong>” of building a super-intelligent AI. Unfortunately, the dialog often jumps over many dangers and lands instead on movie plots like <strong>Wargames</strong> (1983), in which an AI nearly triggers a nuclear war by accidentally misinterpreting human objectives, or <strong>Terminator</strong> (1984), in which an AI weapons system becomes sentient and turns against us with an army of red-eyed robots.</p><p>Both are terrific movies,<strong> but are these the real risks we face</strong>?</p><p>Sure, an accidental nuclear launch or weapons gone rogue are real threats, but military leaders already take those risks seriously. On the other hand, I contend that an <strong>artificial superintelligence</strong> (ASI) could easily subdue humanity without resorting to nukes or killer robots. In fact, it wouldn’t need to use any form of traditional violence. Instead, an ASI would simply <a href="https://bigthink.com/the-future/playing-god-metaverse-mind-control-free-will/">manipulate humanity</a> to meet its own interests.</p><p>I know that sounds like another movie plot, but Big Tech is aggressively developing AI systems with the ability to <strong>influence society at scale</strong>. This isn’t a bug in their design — it’s a goal. That’s because many of the largest corporations deploying AI systems have core business models that involve <em>selling targeted influence</em>. They call it marketing, but when powered by AI to characterize individuals, segment populations, and deploy content aimed at influencing specific groups, it becomes a polarizing force in society.</p><p>We’ve all seen the damage this can cause thanks to <a href="https://bigthink.com/the-present/social-media-distorts-reality/">years of unregulated social media</a> and yet traditional targeting will soon look primitive. In the near future, corporations will be able to target users on an individual-by-individual basis through <a href="https://arxiv.org/abs/2306.11748">personalized interactive conversations.</a> This will be deployed using <strong>artificial agents</strong> that draw us into friendly conversation and casually deploy custom influence that is maximized for our unique views, interests, leanings and backgrounds.</p><p>Now consider this — a few weeks ago Meta unveiled <strong>AI-powered chatbots</strong> on Facebook, Instagram, and WhatsApp through partnerships with “cultural icons and influencers” including Snoop Dogg, Kendall Jenner, Tom Brady, Chris Paul, and Paris Hilton. The goal is to allow users to hold friendly conversations with famous people they admire. We can easily imagine such techniques will soon become a new frontier for targeted influence, deployed as personal conversations with famous faces. It will begin as text chat and voice chat but soon evolve into realistic avatars.</p><p>Why is this so dangerous?</p><p>As I often tell policymakers — think about a skilled salesperson. They know that the best way to sway someone’s opinion is not to hand them a brochure or play them a video. It’s to engage them in conversation, subtly expressing a sales pitch, hearing the customers objections and concerns, and actively working to overcome those barriers. AI systems are now ready to engage individuals this way, performing all steps of the process, and as I detail in this <a href="https://www.iiis.org/DOI2023/HA408FU/">recent academic paper</a> — we humans will be thoroughly outmatched.</p><p>That’s because these AI systems will be far more prepared to target you than any salesperson. They could have access to data about your interests, hobbies, political leanings, personality traits, education, and countless other personal details. And soon, they will be able to read your emotions in your vocal inflections, facial expressions, eye motions and posture.</p><p>You on the other hand, will be talking to an AI that can look like anything from Paris Hilton and Snoop Dogg, to a cute little <a href="https://bigthink.com/the-future/metaverse-fraud-digital-twins/">fairy that guides you through your day</a>. And yet that cute or charming AI could have all the world’s information at its disposal to guide your thinking and counter your objections, while also being trained on sales tactics, cognitive psychology, and strategies of persuasion. And worse, it could just as easily sell you a car as it could convince you of misinformation or propaganda. This risk is called <a href="https://arxiv.org/abs/2306.11748"><em>the AI Manipulation Problem</em></a> and most regulators still fail to appreciate the subtle danger of interactive conversational influence.</p><p>Now let’s look a little further into the future and consider the magnitude of the manipulation risk as AI systems achieve superintelligence. Will we regret that we allowed the largest companies in the world to <a href="https://futureofmarketinginstitute.com/why-generative-ai-is-a-new-form-of-media/">normalize the deployment of AI agents</a> that look human, act human, and sound human, but are NOT human in any real sense of the word, and yet can skillfully pursue human-like tactics that can manipulate our beliefs, influence our opinions, and sway our actions? I think so.</p><p>After all, a superintelligence, by definition, will be an AI system that is significantly smarter than any individual human. And if such an AI system goes rogue, it will not need to take control over our nukes or military drones. It will just need to use the tactics that Big Tech is currently developing — the ability to deploy personalized AI agents that seem so friendly and nonthreatening that we let down our guard, allowing them to <a href="https://medium.com/predict/metaverse-2030-ee59e4d4010d">whisper in our ears</a> while reading our emotions, predicting our actions, and potentially manipulating our behavior with super-human skill.</p><p>This is a real threat and yet we’re not acting like it’s rapidly approaching. In fact, we are deeply <a href="https://venturebeat.com/ai/the-profound-danger-of-conversational-ai/">underestimating the risk</a>, largely because of the personification techniques described above. Today’s AI agents are already so good at pretending to be human, even by text chat, we’re trusting them more than we should. So when these powerful AI systems eventually appear to us as Snoop Dogg or Paris Hilton or some new fictional persona that’s friendly and charming, we will only let down our guard even more.</p><p>So how can we get people to appreciate the magnitude of this risk?</p><p>Over the last decade, I found that an effective way to contextualize the risks is to compare the creation of a superintelligence with the <a href="https://bigthink.com/the-future/general-ai-artificial-intelligence/">arrival of an alien spaceship</a>. I refer to this as the <a href="https://venturebeat.com/ai/snoop-dogg-sentient-ai-and-the-arrival-mind-paradox/"><em>Arrival Mind Paradox</em></a> because the creation of a super-intelligent AI here on earth is arguably more dangerous than intelligent aliens arriving from another planet. And yet with AI now advancing at a record pace, we humans are not acting like we just looked into a telescope and spotted a fleet of ships racing towards us.</p><p>So, let’s compare the relative risks. If an alien spaceship was spotted heading towards earth and moving at a speed that made it five to ten years away, the world would be sharply focused on the approaching entity — hoping for the best, but undoubtedly preparing our defenses, likely in a coordinated global effort unlike anything the world has ever seen. Some would argue that the intelligence species will come in peace, but most would demand that we prepare for a full-scale invasion.</p><p>On the other hand, we have already looked into a telescope that’s pointing back at ourselves and have spotted a <a href="https://www.youtube.com/watch?v=Eu-RyZt_Uas">superintelligence headed for earth</a>. Yes, it will be our own creation, born in a corporate research lab, but it will not be human in any way. And let’s be clear — the fact that we are teaching this intelligence to be good at pretending to be human does not make it less alien. This arriving mind will be profoundly different from us and we have no reason to believe it will possess humanlike values, morals, or sensibilities. And by teaching it to speak our languages , read our emotions, write our programming code and integrate with our computing networks, we are making it a more dangerous threat than an alien from afar.</p><p>Still, we don’t fear the arrival of this alien AI — not in the visceral, stomach turning way that we would fear a mysterious ship headed for earth. That’s the Arrival Mind Paradox — <a href="http://arrivalmind.com/">the fact that we fear the arrival of the wrong aliens</a> and will likely do so until it’s too late to prepare. And if this alien AI shows up looking like Paris Hilton or Snoop Dogg or countless other familiar faces, and speaks to each of us in ways that individually appeal to our personalities and backgrounds, what chance do we have to resist?</p><p>Yes, we should secure our nukes and drones — but we need to have just as aggressive protections against the widespread deployment of personified AI agents. It’s a real threat and we are not prepared.</p><p><strong><em>Louis Rosenberg, PhD</em></strong><em> is a longtime technologist. He is known for founding the early VR company Immersion Corporation (IMMR: Nasdaq) in 1992 and the artificial intelligence company </em><a href="https://unanimous.ai/"><em>Unanimous AI</em></a><em> in 2014. He is also known for developing the first </em><a href="https://spectrum.ieee.org/history-of-augmented-reality"><em>mixed reality system</em></a><em> as a researcher for the U.S. Air Force. He earned his PhD from Stanford and was a professor at California State University.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3958590f50c5" width="1" height="1" alt=""><hr><p><a href="https://medium.com/predict/artificial-superintelligence-and-the-arrival-mind-paradox-3958590f50c5">Artificial Superintelligence and the Arrival Mind Paradox</a> was originally published in <a href="https://medium.com/predict">Predict</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[On Collective Superintelligence]]></title>
            <link>https://medium.com/predict/on-collective-superintelligence-93a4f4f16e4a?source=rss-46e426f3cb75------2</link>
            <guid isPermaLink="false">https://medium.com/p/93a4f4f16e4a</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[chatgpt]]></category>
            <category><![CDATA[government]]></category>
            <category><![CDATA[decision-making]]></category>
            <category><![CDATA[marketing]]></category>
            <dc:creator><![CDATA[Louis Rosenberg, PhD]]></dc:creator>
            <pubDate>Sun, 15 Oct 2023 23:06:53 GMT</pubDate>
            <atom:updated>2023-10-15T23:06:53.581Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ThE3Nm9b3NArS-2VM91p2w.png" /></figure><p><strong>Over the last decade, my goal as an AI researcher</strong> has been to push for the development of AI-powered systems that provide a viable alternative to <em>Artificial Superintelligence —</em> one that achieves super-intelligent capabilities but does so by <em>amplifying human intellect </em>rather than replacing it. My reason is simple: I believe that creating an artificial superintelligence will be an extremely dangerous and dehumanizing step for our species. If you need convincing, see my tiny 2020 picture book entitled <a href="https://www.kirkusreviews.com/book-reviews/louis-rosenberg/arrival-mind/">Arrival Mind</a>.</p><p><strong>That said, is there a safer path to superintelligence?</strong> When I started exploring this back in 2008, I concluded that our best approach would be to create a <a href="https://futurism.com/keeping-humanity-must-cultivate-hive-mind"><strong><em>Collective Superintelligence</em></strong></a> that uses the power of AI to <em>amplify the combined thoughts, opinions, insights, and wisdom </em>of very large groups in real time.<em> </em>No, I’m not talking about using AI to process the <em>combined artifacts</em> of humanity the way LLMs do. Instead, I’m talking about processing the real-time thoughts and opinions of human groups, not to replicate our collective thinking but to facilitate it and amplify it.</p><p><strong>To be honest, I wasn’t quite sure if this was possible</strong> in a seamless and natural way, but <a href="https://arxiv.org/abs/2309.12366">new research</a> at <a href="https://unanimous.ai/">Unanimous AI</a> convinces me this vision is possible. Even more exciting is that the path to Collective Superintelligence will likely include near-term technologies that could transform how large groups work together right now, creating entirely new modes for <strong>Enterprise Collaboration </strong>and <strong>Deliberative Civic Engagement</strong> and even largescale <strong>Governmental Decision-Making</strong>.</p><p>Before I get into any of that, let me give some background:</p><p><strong>My inspiration for Collective Superintelligence (CSi) </strong>has been mother nature. That’s because biological systems have evolved over hundreds of millions of years that show the power of real-time collective intelligence. For example, large fish schools, bee swarms, and bird flocks can think together so effectively, they can function as <a href="https://www.bbc.com/future/article/20161215-why-bees-could-be-the-secret-to-superhuman-intelligence"><em>super-organisms</em></a><em> </em>that seamlessly amplify the collective intelligence of the full group. Biologists call this amplified intellect, <a href="https://www.youtube.com/watch?v=Eu-RyZt_Uas">Swarm Intelligence</a>. The big question I asked over a decade ago was whether human groups could achieve similar results.</p><p><strong>To pursue this goal, I founded Unanimous AI in 2014</strong>. Over the decade since, we have <a href="https://unanimous.ai/publications/">conducted many academic studies </a>that prove human groups can form a Swarm Intelligence that significantly amplifies their collective insights. This has encouraged me that the long-term goal of Collective Superintelligence is viable. We’ve also discovered that it’s a difficult challenge, for it involves understanding the dynamics of human groups as they interact with each other and with real-time AI systems.</p><p><strong>Our core technology is called Artificial Swarm Intelligence </strong>(<a href="https://unanimous.ai/what-is-si/">Swarm AI</a>) and it works well. Famously, I was challenged by a reporter back in 2016 who dared our team 𝘵𝘰 𝘱𝘳𝘦𝘥𝘪𝘤𝘵 𝘵𝘩𝘦 𝘒𝘦𝘯𝘵𝘶𝘤𝘬𝘺 𝘋𝘦𝘳𝘣𝘺 (the first four horses in order) using a “human swarm.” We accepted the challenge, not realizing at the time that it’s considered one of the hardest bets to win in all of sports.</p><p><strong>This is the result as it was reported in Newsweek back then</strong>: <em>“</em><a href="https://www.newsweek.com/artificial-intelligence-turns-20-11000-kentucky-derby-bet-457783"><em>AI turns $20 into $11,000 on Kentucky Derby Bet</em></a><em>.”</em> Of course there was luck involved, but beating 540-to-1 odds was not random chance. Our system (called <a href="https://unanimous.ai/what-is-si/">Swarm</a>) combined the insights of 20 horseracing enthusiasts into a real-time collective intelligence. And it worked:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/941/1*K0DfcrOJwxEW3SPZpY9atA.jpeg" /></figure><p><strong>Also in 2016, we ran the world’s first “Ask me Anything” </strong>with a real-time collective intelligence. We conducted this live on Reddit, which was risky to say the least, but it worked — enabling hundreds of networked people to answer questions as a unified swarm. It drew so much attention, it was chronicled by the <a href="https://www.washingtonpost.com/news/the-intersect/wp/2016/06/02/what-happened-when-an-ai-hive-mind-answered-reddits-burning-politics-questions/">Washington Post in an article</a> where they flagged some of the predictions that came out of this milestone event. In particular, they published a prediction that California would <em>legalize marijuana</em> before the end of the year. Five months later, that prediction <a href="https://time.com/4565438/california-marijuana-faq-rules-prop-64/">came true</a>.</p><p><strong>Over the years since, our Swarm AI technology has been used </strong>by the United Nations, the US Air Force, and Fortune 500 companies. But the grand pursuit of building a <em>Collective Superintelligence</em> seemed far off. That all changed over the last 18 months as new AI methods (including LLMs) have allowed us to quietly pursue a series of breakthroughs that could make a seamless Collective Superintelligence a viable possibility.</p><p><strong>Our most exciting innovation is </strong><a href="https://arxiv.org/abs/2309.12366"><strong>Conversational Swarm Intelligence</strong></a><strong>. </strong>It’s a technology we began pursuing in 2018 but wasn’t viable at scale until we could leverage LLMs in a unique way. It promises to allow groups of almost any size (50 people, 500 people, even 5 million people) to hold real-time conversational deliberations that converge on unified solutions that are amplified by the power of swarm intelligence. The core concept was inspired by fish schools and formal testing shows it actually works.</p><p><strong>How is this related to schools of fish</strong>? Well, it turns out that fish schools can hold real-time “conversations” among thousands of members, rapidly making collective decisions to navigate the ocean with no central authority mediating the process. Each fish communicates with others using a unique organ called a “lateral line” that senses pressure changes in the water caused by neighboring fish. The number of neighbors that a given fish pays attention to varies from species to species, but it’s always a small subset of the group. And because each fish interacts with an <em>overlapping subset</em> of other fish, information quickly propagates across the full population, enabling a unified Swarm Intelligence to emerge.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/675/1*xbqkaHL2KfqsxZhcGnv3bw.png" /><figcaption>Fish school navigating the ocean as a collective super-organism</figcaption></figure><p><strong>Can we enable something similar in humans?</strong> It turns out we can by using a concept we developed in 2018 called <a href="https://ieeexplore.ieee.org/document/9623239"><em>hyperswarms</em></a><em> </em>that divides real-time human groups into overlapping subgroups.<em> </em>For example, we can take a large group of 1000 networked individuals and divide them into 200 groups of 5 people, the members of each subgroup <em>automatically routed</em> into their own unique chat room or videoconferencing room. And if we send them all the same question to discuss or problem to solve, we now have 200 parallel conversations, each reasonably sized for thoughtful deliberations.</p><p><strong>Of course, parallel groups doesn’t create a Swarm Intelligence</strong> because content can’t propagate across the population. We solved this through the novel use of <a href="https://arxiv.org/abs/2306.11748"><em>conversational AI agents</em></a><em> </em>powered by LLMs to emulate the function of the lateral line organ in fish. In particular, we insert an AI agent (referred to as an <a href="https://arxiv.org/ftp/arxiv/papers/2309/2309.12366.pdf">Observer Agent</a>) into each of the 200 subgroups and task that agent with monitoring the dialog in that group, distilling the salient content, and expressing the content in a neighboring group through natural dialog. In this way, each of the 200 groups is given a <em>sixth member</em> that happens to be an AI observer that is tasked with monitoring the real-time insights discussed in one group and conversationally expressing them into one or more other groups, enabling information to propagate across the full population. This can be represented in this diagram:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*mueG-OkF4bFZTDMvmeDtsw.jpeg" /><figcaption>Architecture for building <strong>Collective Superintelligence</strong> (CSi) based on real-time conversations in large groups</figcaption></figure><p><strong>But does this crazy idea work</strong>? To be honest, I wasn’t sure that large groups could hold unified conversations in this way and converge on coherent solutions in real-time — but it turns out, they can. Even more surprising, when have participants compare conversations in a single large chat room with being placed into a conversational swarm, they overwhelmingly prefer the biologically inspired structure. In fact, the vast majority of participants reported feeling like they were much better able to <em>express their views</em> and<em> have their opinions heard</em> in this unique swarm structure than in a traditional structure like a large chat room or forum.</p><p><strong>To explain how Conversational Swarm Intelligence works</strong>, I am sharing a presentation that I gave on October 14th at the 7𝘵𝘩 𝘐𝘯𝘵𝘦𝘳𝘯𝘢𝘵𝘪𝘰𝘯𝘢𝘭 𝘑𝘰𝘪𝘯𝘵 𝘊𝘰𝘯𝘧𝘦𝘳𝘦𝘯𝘤𝘦 𝘰𝘯 𝘈𝘥𝘷𝘢𝘯𝘤𝘦𝘴 𝘪𝘯 𝘊𝘰𝘮𝘱𝘶𝘵𝘢𝘵𝘪𝘰𝘯𝘢𝘭 𝘐𝘯𝘵𝘦𝘭𝘭𝘪𝘨𝘦𝘯𝘤𝘦 (𝘐𝘑𝘊𝘈𝘊𝘐 2023). It describes a research study we conducted earlier this year in collaboration with Carnegie Mellon University. It’s one of many studies we’ve been conducting to assess the benefits of CSI. The results are promising.</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fplayer.vimeo.com%2Fvideo%2F874216544%3Fapp_id%3D122963&amp;dntp=1&amp;display_name=Vimeo&amp;url=https%3A%2F%2Fvimeo.com%2F874216544&amp;image=https%3A%2F%2Fi.vimeocdn.com%2Fvideo%2F1738109409-ea5fc56220320d9519e8c2d650552d9e566c0cb2d884d8b716371aa6b098a1cd-d_1280&amp;key=a19fcc184b9711e1b4764040d3dc5c07&amp;type=text%2Fhtml&amp;schema=vimeo" width="1920" height="1080" frameborder="0" scrolling="no"><a href="https://medium.com/media/0fc0ee2a8aa2e4d570f0fa6a928cc06f/href">https://medium.com/media/0fc0ee2a8aa2e4d570f0fa6a928cc06f/href</a></iframe><p>𝐖𝐡𝐲 𝐢𝐬 𝐭𝐡𝐢𝐬 𝐢𝐦𝐩𝐨𝐫𝐭𝐚𝐧𝐭? Again, I firmly believe we need an alternative to <a href="https://en.wikipedia.org/wiki/Superintelligence"><em>Artificial Superintelligence</em></a> that is constructed using the real-time values, moral, sensibilities, wisdom, and insights of human participants. Again, I’m not talking about processing the artifacts left behind by humans like traditional AI does. I’m talking about <em>connecting human groups in real-time so they can function as a unified superintelligence.</em> If we can achieve this, I believe we can use the power of AI to help humanity make better decisions and solve difficult problems by optimizing our ability to think together.</p><p><strong>If we don’t build a Collective Superintelligence</strong>, I fear we are just building our own digital replacements. For details on why this is a genuine risk, see my <strong>2017 TED talk</strong>: <a href="https://www.youtube.com/watch?v=Eu-RyZt_Uas">https://www.youtube.com/watch?v=Eu-RyZt_Uas</a></p><p><strong>Are you interested in participating in conversational swarms? </strong>Well, if you have a large networked team (50 people or more) that might want to take part in <em>“beta tests”</em> of <em>Conversational Swarm Intelligence</em>, we continue to conduct sessions in collaboration with Carnegie Mellon University. So please, send us an email and ask to find out more: <strong>Gregg@Unanimous.ai</strong></p><p><a href="https://www.linkedin.com/in/louis-rosenberg-025851132/"><strong>Dr. Louis Rosenberg</strong></a><strong> </strong>is an AI researcher and early pioneer of virtual and augmented reality. His work began over 30 years ago in labs at Stanford and NASA. In 1992 he developed the first functional <a href="https://spectrum.ieee.org/history-of-augmented-reality">mixed reality system</a> at Air Force Research Laboratory. In 1993 he founded the early VR company Immersion Corp which he brought public on NASDAQ. In 2004 he founded the early AR company Outland Research. He received his PhD from Stanford, was a professor at California State University, and has been awarded over 300 patents for VR, AR, and AI technologies. He is currently the CEO and Chief Scientist of <a href="https://unanimous.ai/">Unanimous AI</a>.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=93a4f4f16e4a" width="1" height="1" alt=""><hr><p><a href="https://medium.com/predict/on-collective-superintelligence-93a4f4f16e4a">On Collective Superintelligence</a> was originally published in <a href="https://medium.com/predict">Predict</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>