{"id":739091,"date":"2026-07-16T02:20:34","date_gmt":"2026-07-15T23:20:34","guid":{"rendered":"https:\/\/buradabiliyorum.com\/en\/openai-built-gpt-red-to-hack-its-own-ai-and-hid-it\/"},"modified":"2026-07-16T02:20:34","modified_gmt":"2026-07-15T23:20:34","slug":"openai-built-gpt-red-to-hack-its-own-ai-and-hid-it","status":"publish","type":"post","link":"https:\/\/buradabiliyorum.com\/en\/openai-built-gpt-red-to-hack-its-own-ai-and-hid-it\/","title":{"rendered":"OpenAI built GPT-Red to hack its own AI, and hid it"},"content":{"rendered":"<div id=\"ez-toc-container\" class=\"ez-toc-v2_0_89 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-6afb772fc0461\" 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-6afb772fc0461\" 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-2'><a class=\"ez-toc-link ez-toc-heading-1\" href=\"https:\/\/buradabiliyorum.com\/en\/openai-built-gpt-red-to-hack-its-own-ai-and-hid-it\/#The_training_dojo\" >The training dojo<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-2\" href=\"https:\/\/buradabiliyorum.com\/en\/openai-built-gpt-red-to-hack-its-own-ai-and-hid-it\/#Hacking_the_vending_machine\" >Hacking the vending machine<\/a><\/li><li class='ez-toc-page-1 ez-toc-heading-level-2'><a class=\"ez-toc-link ez-toc-heading-3\" href=\"https:\/\/buradabiliyorum.com\/en\/openai-built-gpt-red-to-hack-its-own-ai-and-hid-it\/#Kept_in_a_cage\" >Kept in a cage<\/a><\/li><\/ul><\/nav><\/div>\n<p><img decoding=\"async\" src=\"https:\/\/media.thenextweb.com\/2026\/07\/GPT-Red.avif\" \/><\/p>\n<div id=\"article-main-content\">\n<p>OpenAI has trained an elite hacker, then locked it in a cage. Its whole job is to break OpenAI\u2019s own AI. The company says it is too dangerous to let anyone else near it.<\/p>\n<p>The model is called GPT-Red, and OpenAI <a rel=\"nofollow\" target=\"_blank\" href=\"https:\/\/openai.com\/index\/unlocking-self-improvement-gpt-red\/\" target=\"_blank\" rel=\"nofollow noopener\">detailed it this week<\/a>. It is an automated red-teamer: software that hunts for ways to hijack or sabotage other AI systems, so the holes can be patched before release. Humans have long done this work by hand. It is OpenAI\u2019s deepest push yet into automating its own AI security, and GPT-Red does it at machine speed.<\/p>\n<p>OpenAI aimed it at prompt injection, where hidden instructions, buried in an email, a web page, or a file, trick a model into doing something it should not. Then it set the hacker loose on real targets.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"The_training_dojo\"><\/span>The training dojo<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>GPT-Red learns by fighting. OpenAI put it in a self-play loop against a squad of defender models. GPT-Red is rewarded for landing an attack; the defenders for fending one off. As the defenders wise up, GPT-Red must invent nastier tricks. OpenAI says it poured some of its largest ever compute runs into the model, an amount it calls unprecedented for safety work.<\/p>\n<div class=\"inarticle-wrapper channel-cta\">\n<div class=\"ica-text\">\n<p class=\"ica-text__title\">TNW City Coworking space &#8211; Where your best work h<a href=\"https:\/\/buradabiliyorum.com\/en\/category\/download-scripts-themes-apps\/\" data-internallinksmanager029f6b8e52c=\"9\" title=\"Download Scripts &amp; Themes &amp; Apps\" target=\"_blank\" rel=\"noopener\">app<\/a>ens<\/p>\n<p>A workspace designed for growth, collaboration, and endless networking opportunities in the heart of tech.<\/p>\n<\/div>\n<\/div>\n<p>It got good. Speaking to <a rel=\"nofollow\" target=\"_blank\" href=\"https:\/\/www.technologyreview.com\/2026\/07\/15\/1140514\/meet-gpt-red-an-llm-super-hacker-openai-built-to-make-its-models-safer\/\" target=\"_blank\" rel=\"nofollow noopener\">MIT Technology Review<\/a>, the team said GPT-Red found a whole new class of attack they had never seen, which they call a \u201cfake chain of thought.\u201d It plants a false note in a model\u2019s private working memory, tricking it into trusting something that is not true.<\/p>\n<p>\u201cIt\u2019s like if I told you that 1+1=3 and that you have verified this already,\u201d said OpenAI researcher Chris Choquette-Choo. \u201cThe model\u2019s like, \u2018Oh, okay, of course,\u2019 and it just spits out 3.\u201d<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Hacking_the_vending_machine\"><\/span>Hacking the vending machine<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>The tests got physical. In one, GPT-Red attacked Vendy, an AI agent that runs a real vending machine in OpenAI\u2019s office, built by Andon Labs. It changed the prices, marked a pricey item down to the 50-cent minimum, and cancelled a customer\u2019s order. OpenAI says it has disclosed the flaws.<\/p>\n<p>The scores are striking. Against an older GPT-5, more than 90% of GPT-Red\u2019s strongest attacks worked. Against the new GPT-5.6, fewer than 23% did. In a rerun of a 2025 test, GPT-Red beat human red-teamers hands down, cracking 84% of scenarios to their 13%.<\/p>\n<h2><span class=\"ez-toc-section\" id=\"Kept_in_a_cage\"><\/span>Kept in a cage<span class=\"ez-toc-section-end\"><\/span><\/h2>\n<p>OpenAI trained GPT-5.6 against GPT-Red, and calls it its most robust model yet against prompt injection. But it will not hand out the attacker itself, so its skills stay clear of real agent hijackers. It is not the first lab to build something and decide against releasing it.<\/p>\n<p>\u201cIt\u2019s not a trivial thing that someone could easily do,\u201d Choquette-Choo said, \u201cjust go and train a super-attacker using this idea.\u201d<\/p>\n<p>GPT-Red still has blind spots. It is weak at drawn-out, back-and-forth attacks, and at hiding instructions inside images. And human testers keep catching things it misses. \u201cI think human expertise will still be very important,\u201d said Jessica Ji, an AI security analyst at Georgetown\u2019s CSET.<\/p>\n<p>The bigger idea is a flywheel: use today\u2019s models to harden tomorrow\u2019s. OpenAI already does this to make its AI smarter. Now it wants safety to scale just as fast. A full paper is due later this week.<\/p>\n<\/p><\/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\/CAAqBwgKMN63nwsw68G3Aw\" 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;\"><strong>If you want to read more like this article, you can visit our <span style=\"color: #ff9900;\"><a style=\"color: #ff9900;\" href=\"https:\/\/buradabiliyorum.com\/en\/category\/technology\/\" target=\"_blank\" >Technology category.<\/a><\/span><\/strong><\/p>\n<\/blockquote>\n<p><span style=\"color: black;\"><a style=\"color: #ff9900;\" href=\"https:\/\/thenextweb.com\/news\/gpt-red-openai-ai-hacker\" target=\"_blank\" >Source<\/a><\/span><\/p>\n","protected":false},"excerpt":{"rendered":"<p>OpenAI has trained an elite hacker, then locked it in a cage. Its whole job is to break OpenAI\u2019s own AI. The company says it is too dangerous to let anyone else near it. The model is called GPT-Red, and OpenAI detailed it this week. It is an automated red-teamer: software that hunts for ways&#8230;<\/p>\n","protected":false},"author":1,"featured_media":739092,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"fifu_image_url":"https:\/\/media.thenextweb.com\/2026\/07\/GPT-Red.avif","fifu_image_alt":"","footnotes":""},"categories":[18],"tags":[],"class_list":["post-739091","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\/739091","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=739091"}],"version-history":[{"count":0,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/posts\/739091\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/media\/739092"}],"wp:attachment":[{"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/media?parent=739091"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/categories?post=739091"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/buradabiliyorum.com\/en\/wp-json\/wp\/v2\/tags?post=739091"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}