<?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 Noah Herrin on Medium]]></title>
        <description><![CDATA[Stories by Noah Herrin on Medium]]></description>
        <link>https://medium.com/@noahherrin?source=rss-96938e4b5921------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*vCvx_-rCt6fDh8rg8qJ9oA.jpeg</url>
            <title>Stories by Noah Herrin on Medium</title>
            <link>https://medium.com/@noahherrin?source=rss-96938e4b5921------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 13:27:44 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@noahherrin/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[AI developers will have to overcome our history of social biases]]></title>
            <link>https://medium.com/predict/ai-developers-will-have-to-overcome-our-history-of-social-biases-bae071d574ca?source=rss-96938e4b5921------2</link>
            <guid isPermaLink="false">https://medium.com/p/bae071d574ca</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[ethics]]></category>
            <dc:creator><![CDATA[Noah Herrin]]></dc:creator>
            <pubDate>Sat, 12 Jun 2021 02:06:58 GMT</pubDate>
            <atom:updated>2021-07-14T05:21:10.864Z</atom:updated>
            <content:encoded><![CDATA[<h3>AI Developers will have to Confront Human’s History of Social Biases</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/904/1*9XyA4JQHjfUmpCyTQWx40w.png" /></figure><p>John Lewis was a civil rights leader, member of congress. Yet an Artificial Intelligence system built by Amazon, for facial recognition, incorrectly matched his official congressional photo with a photo of somebody who had been arrested. In a 2018 investigation, the ACLU provided Amazon’s Recognition 25,000 arrest photos, and the photos of every member of congress to see if it could find any matches. The system identified 28 out of 535 congressional representatives as matches to people in arrest photos. It disproportionately had false matches for non-white members of congress, and these are people who lead our country and are recognizable by millions. This is incredibly concerning since some of Amazon’s customers for this service include law and immigration enforcement.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/494/1*qiHRT5qVbt5wUCQs6_Y-_Q.png" /><figcaption>Image Source: <a href="https://www.aclu.org/blog/privacy-technology/surveillance-technologies/amazons-face-recognition-falsely-matched-28">https://www.aclu.org/blog/privacy-technology/surveillance-technologies/amazons-face-recognition-falsely-matched-28</a></figcaption></figure><p>In the next ten years, it’s McKinsey Global Institute has forecasted that Artificial Intelligence will be responsible for adding an additional 13 trillion dollars to the global GDP, an increase of 16 percent by 2030. Artificial Intelligence’s influence impact will come from other industries besides technology, simulations predict that 70 percent of all companies will have adopted some AI technology by 2030. AI can be used to improve the efficiency of respective tasks, meaning the winners of this AI revolution have significant financial incentives. However, it’s critical that everyone knows that these systems are only as good as the information they are trained with. In light of biases against people of color: Academics, Politicians, and Civil Rights groups all agreed that there is a very real concern that these systems will become an automated mechanism for perpetuating historical social biases and discrimination.</p><p>You might be shocked to find out that the Tin Man from 1939’s Wizard of Oz is credited for inspiring philosophers and scientists to ponder the idea of an intelligent machine. In 1950 the father of theoretical computer science, Alan Turing, posed a question to the scientific community.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/426/1*UesmTG6Pv_GJg8cnNgPazg.png" /><figcaption>Image Source: <a href="https://www.under30ceo.com/wp-content/uploads/2014/12/article-2421714-1BD92D16000005DC-452_634x668.jpg">https://www.under30ceo.com/wp-content/uploads/2014/12/article-2421714-1BD92D16000005DC-452_634x668.jpg</a></figcaption></figure><blockquote>Humans use available information as well as reason in order to solve problems and make decisions, so why can’t machines do the same thing?</blockquote><blockquote>-Alan Turing</blockquote><p>Alan Turing’s question set the stage for many of the techniques used today to build artificial intelligence models. Much of the work that goes into developing AI’s is in curating the data that they learn from. Developers of Artificial Intelligence models have to curate large collections of training examples. These examples could be photos for computer vision tasks, text for language processing tasks, or even resumes for an automated hiring system. Each example comes with a label that tells the AI what inference it should draw from the example. Pairs of examples and labels are then given to the AI for training, where the AI uses a series of mathematical techniques to identify relationships within the data that it can associate with different labels. For an example with humans, I ask you to think about what conclusions you’ve been trained to draw when you see somebody in a red hat. Your brain has taken a smaller easy to distinguish feature and associated it with additional meaning.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/484/1*-rbzYTHH9VI3W_SpXR4XZw.png" /><figcaption>Image Source:https://arxiv.org/abs/2005.00813</figcaption></figure><p>Perspective AI is a company that works with Google, to use AI to identify and mitigate toxic dialogue online. Researchers have found that AI has learned an unintended lesson. The AI learned to associate mentions of disability with toxicity This due to the AI noticing that people can be awful and weaponize disabilities as a way to harass on another. Similar sentences varying only by a mention of a disability can receive vastly different toxicity scores. This discovery is very concerning for researchers who worry that AI moderators for social media platforms could accidentally delete and suppress normal discussions about disabilities.</p><p>I spoke to Dr Elglaly, a professor at Western Washington University who is developing a course for software engineers to better understand issues of accessibility in computing. She spoke about a chatbot that Microsoft developed years ago. The chatbot was first released in China, where it was massively successful. Later, Microsoft made a twitter account for the same chatbot, and <a href="https://www.theverge.com/2016/3/24/11297050/tay-microsoft-chatbot-racist">the twitter community radicalized it and it had to get taken down</a>, not before it had attacked feminism and praised the Nazi’s. Dr. Elglaly commented, “Despite the technology being successful in one environment, and when the same technology was moved into another environment it had a different outcome… so It’s not necessarily the fault of the technology, the environment the key difference“. Dr. Elglaly emphasized that engineers need to do more throughout the entire lifecycle of an AI to ensure that its functionality is maintained across the many environments it’s used in.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/780/0*HxVDKU_DvgcbI80t.jpg" /><figcaption>Image Source: <a href="https://www.google.com/url?sa=i&amp;url=https%3A%2F%2Fwww.seattletimes.com%2Fbusiness%2Famazon%2Fdespite-hq2-plan-amazons-seattle-real-estate-target-grows-to-14-million-square-feet%2F&amp;psig=AOvVaw2AVgyN7-ZVbDGB9sYwX7ra&amp;ust=1623549896226000&amp;source=images&amp;cd=vfe&amp;ved=0CAIQjRxqFwoTCJDZ_emAkfECFQAAAAAdAAAAABAD">https://www.google.com/url?sa=i&amp;url=https%3A%2F%2Fwww.seattletimes.com%2Fbusiness%2Famazon%2Fdespite-hq2-plan-amazons-seattle-real-estate-target-grows-to-14-million-square-feet%2F&amp;psig=AOvVaw2AVgyN7-ZVbDGB9sYwX7ra&amp;ust=1623549896226000&amp;source=images&amp;cd=vfe&amp;ved=0CAIQjRxqFwoTCJDZ_emAkfECFQAAAAAdAAAAABAD</a></figcaption></figure><p>Developers of AI need to take action to ensure that the data that is used to train AI is representative of the environment it will be used in. Dr. Elglaly spoke about a controversy surrounding a resume reviewing AI, which studied 10 years of resumes that amazon had collected and the hiring outcomes of those resumes. Of those resumes, the majority of them came from male applicants, likely due to technology being a male dominated industry. Since the majority of the hired employees were men, it negatively associated the word “Women’s” on resumes as something to penalize. Applicant’s with “Women’s Computing Club” or “Women’s College” in their resumes would see their resumes ranked lower than similar resumes without the word “Women’s”. If engineers do not intervene, the data will define the environment that an AI will work in, leading to the perpetuation of inequal treatment especially in the case of gender biases.</p><p>The idea of having an AI maneuver through traffic without my attention is an exciting prospect, especially in my home city of Seattle. The safety of autonomous vehicles both for the passengers and pedestrians has been a prominent focus for the media, however, not everyone is being considered as stakeholder. Jutta Treviranus, the director for the inclusive design research center at the Ontario College of Art &amp; Design, presented various tested AI models responsible for piloting autonomous vehicles with photos of a friend in a wheelchair about to cross the road. All the models chose to run her friend over. The engineers working on the models said that they were immature and asked her to return when they had improved the models, when she tested the improved models they ran over her friend with greater confidence.</p><p>Despite AI’s awe-inspiring potential to define the future of humankind, it’s important for everyone to be aware of its limitations especially in terms of the data it’s given. It’s impossible to foresee every possible scenario, but if we ensure that everyone is included as a critic and a designer then we can help to ensure that AI is built with all of us in mind.</p><p><strong>Sources</strong></p><p>Article: <em>Notes from the AI frontier: Modeling the impact of AI on the world economy </em>[<a href="https://www.mckinsey.com/featured-insights/artificial-intelligence/notes-from-the-ai-frontier-modeling-the-impact-of-ai-on-the-world-economy#part1">Link</a>]</p><p>Article: Amazon scraps secret AI recruiting tool that showed bias against women [<a href="https://www.reuters.com/article/us-amazon-com-jobs-automation-insight/amazon-scraps-secret-ai-recruiting-tool-that-showed-bias-against-women-idUSKCN1MK08G">Link</a>]</p><p>Article: <em>Amazon’s Face Recognition Falsely Matched 28 Members of Congress With Mugshots </em>[<a href="https://www.aclu.org/blog/privacy-technology/surveillance-technologies/amazons-face-recognition-falsely-matched-28">Link</a>]</p><p>Research Paper: <em>Social Biases in NLP Models as Barriers for Persons with Disabilities </em>[<a href="https://arxiv.org/abs/2005.00813">Link</a>]</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=bae071d574ca" width="1" height="1" alt=""><hr><p><a href="https://medium.com/predict/ai-developers-will-have-to-overcome-our-history-of-social-biases-bae071d574ca">AI developers will have to overcome our history of social biases</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 pick a job that won’t get stolen?]]></title>
            <link>https://noahherrin.medium.com/how-to-pick-a-job-that-wont-get-stolen-3689da2f7d6e?source=rss-96938e4b5921------2</link>
            <guid isPermaLink="false">https://medium.com/p/3689da2f7d6e</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[wwu]]></category>
            <dc:creator><![CDATA[Noah Herrin]]></dc:creator>
            <pubDate>Wed, 21 Feb 2018 17:48:30 GMT</pubDate>
            <atom:updated>2018-02-21T17:48:30.295Z</atom:updated>
            <content:encoded><![CDATA[<p>“If you know the enemy and know yourself, you need not fear the result of a hundred battles. if you know yourselves but not the enemy, for every victory gained you will also suffer a defeat. If you now neither the enemy nor yourself, you will succumb in every battle.”</p><p>-Sun Tzu, The Art of War</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qXHBL9W-VSRILhYMmXpQ2Q.jpeg" /><figcaption><a href="https://www.pinterest.com/pin/278238083208881941/">https://www.pinterest.com/pin/278238083208881941/</a></figcaption></figure><p>While this may not be the conventional war that Sun Tzu wrote about in his book, The Art of War, This battle over the future of our workforce is no less important. Unless people and institutions take steps to prepare for the automated world, hundreds of millions of jobs will be displaced or lost in the coming decades.</p><p>Before we talk about the specific things you can do, lets cover the basic stuff you should know. According to a <a href="https://obamawhitehouse.archives.gov/sites/whitehouse.gov/files/documents/Artificial-Intelligence-Automation-Economy.PDF">report</a> released by the White House there is a strong connection between job security and job training. Keep that in mind when you’re grinding out an essay at 2 am. that degree you’re working for will make a big difference.</p><p>First lets explore the strengths of our enemy, the machines.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ikf7XLdoAllQuO2kfZRPhg.jpeg" /></figure><p>Strengths of Automation</p><ul><li>Able to quickly and accurately complete procedural tasks.</li><li>Can analyze large quantities of data using artificial intelligence.</li><li>Can complete complex calculations easily.</li><li>Consistent (Won’t call in sick).</li><li>Able to worker longer hours.</li></ul><p>The main strength of the machines is their consistency, they function the same throughout the day and will preform near perfectly so long as the situation remains the same. Knowing this, you need to understand that humans won’t be able to accomplish tasks meant for robots faster. So don’t be faster, be smarter!</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*MCr38XbkbVfCVCwF3Ovl3w.jpeg" /><figcaption><a href="https://www.ocregister.com/2016/03/17/robots-are-taking-our-jobs-and-thats-a-good-thing/">https://www.ocregister.com/2016/03/17/robots-are-taking-our-jobs-and-thats-a-good-thing/</a></figcaption></figure><p>Strengths of Humans</p><ul><li>Ability to communicate non quantitative ideas.</li><li>Creativity, can come up with new ideas.</li><li>Problem solving.</li><li>Being human, there are many luxury services that would lose what makes them luxury if a robot were to be preforming them.</li></ul><p>Did you notice that almost all of the strengths of human workers are things that aren’t typically taught in school. After all if you can teach a human you can also teach a robot.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/768/1*EWG9zCZa46njNatVFk1Ufg.jpeg" /><figcaption><a href="https://www.nytimes.com/2017/12/11/style/robots-jobs-children.html">https://www.nytimes.com/2017/12/11/style/robots-jobs-children.html</a></figcaption></figure><p>I know what you’re thinking. Picking a career is hard and I need inspiration. As a college student I completely understand that feeling, so I decided to offer a few ideas that experts from the New York Post and Bloomberg have decided will either stand the test of time, or become obsolete.</p><p>What are some Jobs are about to go?</p><ul><li>Cashiers</li><li>Drivers</li><li>Retail Workers</li><li>Food Service Workers</li></ul><p>What jobs are here to stay?</p><ul><li>Police Officer</li><li>Management</li><li>Healthcare professionals</li><li>and of’course Beyonce</li></ul><figure><img alt="" src="https://cdn-images-1.medium.com/max/900/0*dfi1Cv9rCzLl5QT9.jpg" /><figcaption><a href="https://www.radyofenomen.com/muzik-haber/beyonce-nin-basari-sirri.html">https://www.radyofenomen.com/muzik-haber/beyonce-nin-basari-sirri.html</a></figcaption></figure><ul><li>All jokes aside, many jobs in the entertainment industry will also be safe</li></ul><p>While this isn’t the bible of surviving the automation apocalypse it contains valuable information that will help you prepare for the coming storm. Remember that robots are machines that can and will do whatever they are told to do with super human speed and precision. It’s not worth fighting an uphill battle when you could take advantage of your strengths. Humans are creative and we work best together because robots and artificial intelligence can’t compete with our communication skills. if you can take anything away from this article let it be remember that what makes you a valuable member of the workforce is the same stuff that makes you human.</p><p>That and don’t be a driver, you don’t want to see Elon Musk’s smug smirk when his cars steal your job.</p><p>If you’ve already entered the workforce or have decided on what your future career is, let me know how you think it will fair against automation.</p><p><a href="http://www.strawpoll.me/15085886">How worried about your job being automated?</a></p><p>Works Cited</p><p>Gaskell, Adi. “What Factors Might Influence The Automation Of Jobs?” <em>Forbes</em>, Forbes Magazine, 15 Feb. 2018, <a href="http://www.forbes.com/sites/adigaskell/2018/02/15/what-factors-might-influence-the-automation-of-jobs/#5f70f1143dad.">www.forbes.com/sites/adigaskell/2018/02/15/what-factors-might-influence-the-automation-of-jobs/#5f70f1143dad.</a></p><p>Green, Jeff. “Automation Creates Greater Risk for Women’s Jobs, Davos Report Finds.” <em>Los Angeles Times</em>, Los Angeles Times, 28 Jan. 2018, www.latimes.com/business/la-fi-women-robots-jobs-20180128-story.html.</p><p>Lee, Kristin. “Artificial Intelligence, Automation, and the Economy.” <em>National Archives and Records Administration</em>, National Archives and Records Administration, 20 Dec. 2016, obamawhitehouse.archives.gov/blog/2016/12/20/artificial-intelligence-automation-and-economy.</p><p>RobotWorx. “Advantages of Automation.” <em>Industrial Robots</em>, RobotWorx, 6 May 2014, <a href="http://www.robots.com/articles/viewing/advantages-of-automation.">www.robots.com/articles/viewing/advantages-of-automation.</a></p><p>Topping, Amanda. “Are Security Jobs Safe from Automation?” <em>Total Security Daily Advisor</em>, 12 Feb. 2018, totalsecuritydailyadvisor.blr.com/emerging-issues-in-security/security-jobs-safe-automation/.</p><p>Tucker, Reed. “7 Jobs Robots Can’t Steal from You.” <em>New York Post</em>, New York Post, 26 June 2017, nypost.com/2017/06/24/7-jobs-robots-cant-steal-from-you/.</p><p>Whitehouse, Marc, et al. “Is Your Job About to Disappear.” <em>Bloomberg</em>, 22 June 2017, <a href="http://www.bloomberg.com/graphics/2017-jobs-automation-risk/">www.bloomberg.com/graphics/2017-jobs-automation-risk/.</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=3689da2f7d6e" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>