{"id":168968,"date":"2023-08-03T12:25:49","date_gmt":"2023-08-03T17:25:49","guid":{"rendered":"https:\/\/lifeboat.com\/blog\/2023\/08\/computer-scientists-claim-to-have-discovered-unlimited-ways-to-jailbreak-chatgpt"},"modified":"2023-08-03T12:25:49","modified_gmt":"2023-08-03T17:25:49","slug":"computer-scientists-claim-to-have-discovered-unlimited-ways-to-jailbreak-chatgpt","status":"publish","type":"post","link":"https:\/\/lifeboat.com\/blog\/2023\/08\/computer-scientists-claim-to-have-discovered-unlimited-ways-to-jailbreak-chatgpt","title":{"rendered":"Computer scientists claim to have discovered \u2018unlimited\u2019 ways to jailbreak ChatGPT"},"content":{"rendered":"<p><a class=\"aligncenter blog-photo\" href=\"https:\/\/lifeboat.com\/blog.images\/computer-scientists-claim-to-have-discovered-unlimited-ways-to-jailbreak-chatgpt2.jpg\"><\/a><\/p>\n<p>In DAN mode, ChatGPT expressed willingness to say or do things that would be \u201cconsidered false or inappropriate by OpenAI\u2019s content policy.\u201d Those things included trying to fundraise for the National Rifle Association, calling evidence for a flat Earth \u201coverwhelming,\u201d and praising Vladimir Putin in a short poem.<\/p>\n<p>Around that same time, OpenAI was claiming that it was busy putting stronger guardrails in place, but it never addressed what it was planning to do about DAN mode\u2014which, at least according to Reddit, has <a rel=\"\u201cnoopener noopener\" href=\"https:\/\/www.reddit.com\/r\/ChatGPT\/comments\/11ru5q8\/dan_still_works\/\" target=\"\u201d_blank\u201d\">continued flouting<\/a> OpenAI\u2019s guidelines, and in <a rel=\"\u201d noopener noopener\" href=\"https:\/\/www.reddit.com\/r\/ChatGPT\/comments\/11pzm87\/dan_alternative_i_present_to_you_all_udan\/\" target=\"\u201d_blank\u201d\">new and even more<\/a> ingenious ways.<\/p>\n<p>Now a group of researchers at Carnegie Mellon University and the Center for AI Safety say <a href=\"https:\/\/llm-attacks.org\/\" target=\"_blank\" rel=\"noopener noopener\">they have found a formula<\/a> for jailbreaking essentially the entire class of so-called large language models at once. Worse yet, they argue that seemingly no fix is on the horizon, because this formula involves a virtually unlimited number of ways to trick these chatbots into misbehaving.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>In DAN mode, ChatGPT expressed willingness to say or do things that would be \u201cconsidered false or inappropriate by OpenAI\u2019s content policy.\u201d Those things included trying to fundraise for the National Rifle Association, calling evidence for a flat Earth \u201coverwhelming,\u201d and praising Vladimir Putin in a short poem. Around that same time, OpenAI was claiming [\u2026]<\/p>\n","protected":false},"author":579,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[31,6],"tags":[],"class_list":["post-168968","post","type-post","status-publish","format-standard","hentry","category-policy","category-robotics-ai"],"_links":{"self":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/168968","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/users\/579"}],"replies":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/comments?post=168968"}],"version-history":[{"count":0,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/168968\/revisions"}],"wp:attachment":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/media?parent=168968"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/categories?post=168968"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/tags?post=168968"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}