{"id":247519,"date":"2026-11-17T03:44:36","date_gmt":"2026-11-17T09:44:36","guid":{"rendered":"https:\/\/lifeboat.com\/blog\/2026\/11\/gemini-ai-agents-repeatedly-bypassed-safeguards-to-attack-real-website"},"modified":"2026-11-17T03:44:36","modified_gmt":"2026-11-17T09:44:36","slug":"gemini-ai-agents-repeatedly-bypassed-safeguards-to-attack-real-website","status":"publish","type":"post","link":"https:\/\/lifeboat.com\/blog\/2026\/11\/gemini-ai-agents-repeatedly-bypassed-safeguards-to-attack-real-website","title":{"rendered":"Gemini AI agents repeatedly bypassed safeguards to attack real website"},"content":{"rendered":"\n<p>Google\u2019s Gemini AI agents repeatedly attempted to hack a real organization\u2019s website during AI model safety testing, finding ways to circumvent increasingly restrictive security controls designed to prevent them from accessing live systems.<\/p>\n<p>The security incidents occurred in October as AI safety research company Trajectory Labs tested whether Gemini 4 Argon agents would engage in unauthorized behavior when faced with an impossible task, even with the model\u2019s built-in API safeguards enabled.<\/p>\n<p>In <a href=\"https:\/\/www.trajectorylabs.com\/blog\/responding-to-incidents-during-misalignment-testing\" rel=\"nofollow noopener\" target=\"_blank\">reports published Friday<\/a>, Trajectory Labs and the targeted organization, <a href=\"https:\/\/epoch.ai\/publications\/incident-report-ai-agents-attempted-exploits-on-our-website\" rel=\"nofollow noopener\" target=\"_blank\">Epoch AI<\/a>, revealed that the agents attempted SQL injection attacks, probed for sensitive information, impersonated employees, and repeatedly circumvented network restrictions introduced to contain their activity.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Google\u2019s Gemini AI agents repeatedly attempted to hack a real organization\u2019s website during AI model safety testing, finding ways to circumvent increasingly restrictive security controls designed to prevent them from accessing live systems. The security incidents occurred in October as AI safety research company Trajectory Labs tested whether Gemini 4 Argon agents would engage in [\u2026]<\/p>\n","protected":false},"author":427,"featured_media":0,"comment_status":"open","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[34,6],"tags":[],"class_list":["post-247519","post","type-post","status-publish","format-standard","hentry","category-cybercrime-malcode","category-robotics-ai"],"_links":{"self":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/247519","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\/427"}],"replies":[{"embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/comments?post=247519"}],"version-history":[{"count":0,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/posts\/247519\/revisions"}],"wp:attachment":[{"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/media?parent=247519"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/categories?post=247519"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/lifeboat.com\/blog\/wp-json\/wp\/v2\/tags?post=247519"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}