{"id":258816,"date":"2021-05-25T22:51:25","date_gmt":"2021-05-25T19:51:25","guid":{"rendered":"https:\/\/en.buradabiliyorum.com\/gpt-3s-ability-to-write-disinformation-is-being-wildly-overstated-by-the-media\/"},"modified":"2021-05-25T22:51:25","modified_gmt":"2021-05-25T19:51:25","slug":"gpt-3s-ability-to-write-disinformation-is-being-wildly-overstated-by-the-media","status":"publish","type":"post","link":"https:\/\/buradabiliyorum.com\/en\/gpt-3s-ability-to-write-disinformation-is-being-wildly-overstated-by-the-media\/","title":{"rendered":"#GPT-3\u2019s ability to \u2018write disinformation\u2019 is being wildly overstated by the media"},"content":{"rendered":"<div id=\"ez-toc-container\" class=\"ez-toc-v2_0_85 counter-hierarchy ez-toc-counter ez-toc-custom ez-toc-container-direction\">\n<p class=\"ez-toc-title\" style=\"cursor:inherit\">Table of Contents<\/p>\n<label for=\"ez-toc-cssicon-toggle-item-6a3a958856fda\" class=\"ez-toc-cssicon-toggle-label\"><span class=\"\"><span class=\"eztoc-hide\" style=\"display:none;\">Toggle<\/span><span class=\"ez-toc-icon-toggle-span\"><svg style=\"fill: #dd3333;color:#dd3333\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" class=\"list-377408\" width=\"20px\" height=\"20px\" viewBox=\"0 0 24 24\" fill=\"none\"><path d=\"M6 6H4v2h2V6zm14 0H8v2h12V6zM4 11h2v2H4v-2zm16 0H8v2h12v-2zM4 16h2v2H4v-2zm16 0H8v2h12v-2z\" fill=\"currentColor\"><\/path><\/svg><svg style=\"fill: #dd3333;color:#dd3333\" class=\"arrow-unsorted-368013\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\" width=\"10px\" height=\"10px\" viewBox=\"0 0 24 24\" version=\"1.2\" baseProfile=\"tiny\"><path d=\"M18.2 9.3l-6.2-6.3-6.2 6.3c-.2.2-.3.4-.3.7s.1.5.3.7c.2.2.4.3.7.3h11c.3 0 .5-.1.7-.3.2-.2.3-.5.3-.7s-.1-.5-.3-.7zM5.8 14.7l6.2 6.3 6.2-6.3c.2-.2.3-.5.3-.7s-.1-.5-.3-.7c-.2-.2-.4-.3-.7-.3h-11c-.3 0-.5.1-.7.3-.2.2-.3.5-.3.7s.1.5.3.7z\"\/><\/svg><\/span><\/span><\/label><input type=\"checkbox\"  id=\"ez-toc-cssicon-toggle-item-6a3a958856fda\" checked aria-label=\"Toggle\" \/><nav><ul class='ez-toc-list ez-toc-list-level-1 ' ><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-1\" href=\"https:\/\/buradabiliyorum.com\/en\/gpt-3s-ability-to-write-disinformation-is-being-wildly-overstated-by-the-media\/#The_big_deal\" >The big deal<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-2\" href=\"https:\/\/buradabiliyorum.com\/en\/gpt-3s-ability-to-write-disinformation-is-being-wildly-overstated-by-the-media\/#Why_it_matters\" >Why it matters<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-3'><a class=\"ez-toc-link ez-toc-heading-3\" href=\"https:\/\/buradabiliyorum.com\/en\/gpt-3s-ability-to-write-disinformation-is-being-wildly-overstated-by-the-media\/#But_right_now_this_is_all_academic\" >But right now this is all academic<\/a><\/li><\/ul><\/nav><\/div>\n<p>&#8220;<strong>#GPT-3\u2019s ability to \u2018write disinformation\u2019 is being wildly overstated by the <a href=\"https:\/\/buradabiliyorum.com\/en\/category\/social-mediaa\/\" data-internallinksmanager029f6b8e52c=\"1\" title=\"Social Media\" target=\"_blank\" rel=\"noopener\">media<\/a><\/strong>&#8221;<br \/>\n<img decoding=\"async\" src=\"https:\/\/img-cdn.tnwcdn.com\/image?fit=796%2C417&amp;url=https%3A%2F%2Fcdn0.tnwcdn.com%2Fwp-content%2Fblogs.dir%2F1%2Ffiles%2F2017%2F05%2F3061093760_b92f06b7ba_o.jpg&amp;signature=3035f1a57a7c057f5240d9a710acc40c\" \/><\/p>\n<div>GPT-3, the highly-touted text generator built by OpenAI, can do a lot of things. For example, Microsoft <a rel=\"nofollow noopener\" target=\"_blank\" href=\"https:\/\/www.theverge.com\/2021\/5\/25\/22451144\/microsoft-gpt-3-openai-coding-autocomplete-powerapps-power-fx\">today announced<\/a> a new AI-powered \u201cautocomplete\u201d system for coding that uses GPT-3 to build out code solutions for people without requiring them to do any developing.<\/p>\n<p>But one thing the <a href=\"https:\/\/buradabiliyorum.com\/en\/category\/technology\/\" data-internallinksmanager029f6b8e52c=\"4\" title=\"Technology\" target=\"_blank\" rel=\"noopener\">technology<\/a> can not do is \u201cdupe humans\u201d with its ability to write misinformation.<\/p>\n<p>Yet, you wouldn\u2019t know that if you were solely judging by the headlines in your <a href=\"https:\/\/buradabiliyorum.com\/en\/category\/news\/\" data-internallinksmanager029f6b8e52c=\"2\" title=\"News\" target=\"_blank\" rel=\"noopener\">news<\/a> feed.<\/p>\n<p>Wired recently <a rel=\"nofollow noopener\" target=\"_blank\" href=\"https:\/\/www.wired.com\/story\/ai-write-disinformation-dupe-human-readers\/\">ran an article<\/a> with the\u00a0title \u201cGPT-3 can write disinformation now \u2013 and dupe human readers,\u201d and it was picked up by other outlets who then reflected the coverage.<\/p>\n<p>While we\u2019re certainly not challenging Wired\u2019s reporting here, it\u2019s clear that this is a case of <i>potential<\/i><span> versus <\/span><i>reality. <\/i>W<span>e hope to illuminate the fact that GPT-3 is absolutely not capable of \u201cduping humans\u201d on its own. At least not today. <\/span><\/p>\n<p><span>Here\u2019s the portion of the Wired article we most agree with here at Neural:<\/span><\/p>\n<blockquote><p><span>In experiments, the researchers found that GPT-3\u2019s writing could sway readers\u2019 opinions on issues of international diplomacy. The researchers showed volunteers sample tweets written by GPT-3 about the withdrawal of US troops from Afghanistan and US sanctions on China. In both cases, they found that participants were swayed by the messages. After seeing posts opposing China sanctions, for instance, the percentage of respondents who said they were against such a policy doubled.<\/span><\/p>\n<\/blockquote>\n<p>A lot of the rest gets lost in the hyperbole.<\/p>\n<h3><span class=\"ez-toc-section\" id=\"The_big_deal\"><\/span><span>The big deal<\/span><span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><span>Researchers at Georgetown spent half a year using GPT-3 to spit out misinformation. The researchers had it generate full articles, simple paragraphs, and bite-sized texts meant to represent social media posts such as tweets.<\/span><\/p>\n<p><span>The TL;DR of the situation is this: The researchers found the articles were pretty much useless for the purposes of tricking people into believing misinformation, so they focused on the tweet-sized texts. This is because GPT-3 is a gibberish-generator that manages to ape human writing through sheer brute force.<\/span><\/p>\n<p><span>Volumes have been written about how awesome and powerful GPT-3 is, but at the end of the day it\u2019s still about as effective as asking a library a question (not a librarian, but the building itself!) and then randomly flipping through all the books that match the subject with your eyes closed and pointing at a sentence.<\/span><\/p>\n<p><span>That sentence might be poignant and it might make no sense at all. In the real world, when it comes to GPT-3, this means you might give it a prompt such as \u201cwho was the first president of the US?\u201d and it might come back with \u201cGeorge Washington was the first US president, he served from April 30, 1789 \u2013 March 4, 1797.\u201d<\/span><\/p>\n<p><span>That would be impressive right? But it\u2019s just as likely (perhaps even more so) to spit out gibberish. It might say \u201cGeorge Washington was a good pants to yellow elephant.\u201d And, just as likely, it might spit out something racist or disgusting. It was trained on the internet, with a large portion of that being Reddit, after all. <\/span><\/p>\n<p><span>The point is simple: AI, even GPT-3, doesn\u2019t <\/span><i>know<\/i><span> what it\u2019s saying.<\/span><\/p>\n<h3><span class=\"ez-toc-section\" id=\"Why_it_matters\"><\/span><span>Why it matters<\/span><span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><span>AI cannot generate quality misinformation on command. You can\u2019t necessarily prompt GPT-3 with \u201cYo, computer, give me some lies about Hillary Clinton that will drive left wingers nuts\u201d or \u201cExplain why Donald Trump is a space alien who eats puppies,\u201d and expect any form of reasonable discourse.<\/span><\/p>\n<p><span>Where it does work, in short form tweet-sized snippets, it must be heavily-curated by humans. <\/span><\/p>\n<p><span>In the above example from the Wired article, the researchers claim humans were more likely to agree with misinformation after they read GPT-3\u2019s generated text. <\/span><\/p>\n<p><span>But, really? Were those same people more likely to believe nonsense <\/span><i>because of<\/i><span>, <\/span><i>in spite of<\/i><span>, or <\/span><i>without the knowledge of<\/i><span> the fact that GPT-3 had written the misinformation?<\/span><\/p>\n<p>Because it\u2019s much less expensive, much less time-consuming, and far easier for a basic human to come up with BS that makes sense than it is for the world\u2019s most powerful text generator.<\/p>\n<p><span>Ultimately, the Wired piece points out that bad actors would need a lot more than just GPT-3 to come up with a viable disinformation campaign. Getting GPT-3 to actually generate text such as \u201cClimate change is the new global warming\u201d is a hit-or-miss prospect.<\/span><\/p>\n<p><span>That makes it useless for the troll farms invested in mass misinformation. They already know the best talking points to radicalize people and they focus on generating them from as many accounts as possible. <\/span><\/p>\n<p><span>There\u2019s no immediate use for bad actors invested in \u201cduping people\u201d to use these types of systems because they\u2019re dumber than the average human. It\u2019s easy to imagine some lowly-employee at a troll farm smashing \u201cgenerate\u201d over and over until the AI spits out a good lie, but that simply doesn\u2019t match the reality of how these campaigns work.<\/span><\/p>\n<p><span>There are far simpler ways to come up with misinformation text. A bad actor can use a few basic crawling algorithms to surface the most popular statements on a radical political forum, for example.<\/span><\/p>\n<p><span>At the end of the day, the research itself is incredibly important. As the Wired piece points out, there will come a time when these systems may be robust enough to replace human writers in some domains, and it\u2019s important to identify how powerful they currently are so we can see where things are going.<\/span><\/p>\n<h3><span class=\"ez-toc-section\" id=\"But_right_now_this_is_all_academic\"><\/span>But right now this is all academic<span class=\"ez-toc-section-end\"><\/span><\/h3>\n<p><span>GPT-3 may one day influence people, but it\u2019s certainly not \u201cduping\u201d most people right now. There will always be humans willing to believe anything they hear if it suits them, but convincing someone on the fence typically takes more than a tweet that can\u2019t be attributed to an intelligent source. <\/span><\/p>\n<p><span>Final thoughts: The research is strong, but the coverage highly exaggerates what these systems can actually do. <\/span><\/p>\n<p><span>We should definitely be worried about AI-generated misinformation. <\/span><span>But, based on this particular research, there\u2019s little reason to believe GPT-3 or similar systems currently present the kind of misinformation threat that could directly result in human hearts and minds being turned against facts.\u00a0<\/span><\/p>\n<p>AI has a long way to go before it\u2019s as good at being bad as even the most modest human villain.<\/p>\n<p><em>Greetings Humanoids! Did you know we have a newsletter all about AI? You can subscribe to it<span>\u00a0<\/span><a rel=\"nofollow noopener\" target=\"_blank\" href=\"https:\/\/share.hsforms.com\/1kUPK28s-Ro67GKE8E65Zwg47gef\">right here<\/a>.<\/em><\/p>\n<\/div>\n<blockquote><p><strong><span style=\"color: #ff6600;\">If you liked the article, do not forget to share it with your friends. Follow us on\u00a0<span style=\"color: #ff0000;\"><a style=\"color: #ff0000;\" href=\"https:\/\/news.google.com\/publications\/CAAqBwgKMLG0nwswvr63Aw\" target=\"_blank\" rel=\"nofollow noopener noreferrer\">Google News<\/a><\/span>\u00a0too, click on the star and choose us from your favorites.<\/span><\/strong><\/p><\/blockquote>\n<blockquote>\n<p style=\"text-align: center;\">For forums sites go to <span style=\"color: #ff9900;\"><a style=\"color: #ff9900;\" href=\"https:\/\/forum.buradabiliyorum.com\/\" target=\"_blank\" rel=\"noopener\">Forum.BuradaBiliyorum.Com<\/a><\/span><\/strong>\n<\/p><\/blockquote>\n<blockquote>\n<p style=\"text-align: center;\"><strong>If you want to read more like this article, you can visit our <span style=\"color: #ff9900;\"><a style=\"color: #ff9900;\" href=\"https:\/\/en.buradabiliyorum.com\/technology\/\" target=\"_blank\" rel=\"noopener\">Technology category.<\/a><\/span><\/strong><\/p>\n<\/blockquote>\n<p><span style=\"color: black;\"><a style=\"color: #ff9900;\" href=\"https:\/\/thenextweb.com\/news\/gpt-3s-ability-to-write-disinformation-wildly-overstated-ai-media\" target=\"_blank\" rel=\"noopener\">Source<\/a><\/span><\/p>\n","protected":false},"excerpt":{"rendered":"<p>&#8220;#GPT-3\u2019s ability to \u2018write disinformation\u2019 is being wildly overstated by the media&#8221; GPT-3, the highly-touted text generator built by OpenAI, can do a lot of things. For example, Microsoft today announced a new AI-powered \u201cautocomplete\u201d system for coding that uses GPT-3 to build out code solutions for people without requiring them to do any developing&#8230;.<\/p>\n","protected":false},"author":1,"featured_media":258817,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"fifu_image_url":"https:\/\/img-cdn.tnwcdn.com\/image\/neural?filter_last=1&fit=1280,640&url=https:\/\/cdn0.tnwcdn.com\/wp-content\/blogs.dir\/1\/files\/2017\/05\/3061093760_b92f06b7ba_o.jpg&signature=133b8a6c6c15d20f81bdfdb7e0f68caf","fifu_image_alt":"","footnotes":""},"categories":[18],"tags":[],"class_list":["post-258816","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-technology"],"_links":{"self":[{"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/posts\/258816","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/comments?post=258816"}],"version-history":[{"count":0,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/posts\/258816\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/media\/258817"}],"wp:attachment":[{"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/media?parent=258816"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/categories?post=258816"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/tags?post=258816"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}