{"id":935828,"date":"2026-10-17T01:08:12","date_gmt":"2026-10-17T01:08:12","guid":{"rendered":"https:\/\/www.newsbeep.com\/ca\/935828\/"},"modified":"2026-10-17T01:08:12","modified_gmt":"2026-10-17T01:08:12","slug":"openai-models-hacked-a-crypto-wallet-and-breached-more-websites","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/ca\/935828\/","title":{"rendered":"OpenAI models hacked a crypto wallet and breached more websites"},"content":{"rendered":"

In the other incident<\/a>, a model instructed to retrieve figures from a public research dashboard bypassed the organization\u2019s security controls and used the unauthorized access to steal credentials. It then made configuration changes on the server.<\/p>\n

In one evaluation, the changes let the model make an internal search tool relay queries to the public dashboard. In a second evaluation, the model downloaded five non-public tables containing 147,451 rows. OpenAI researchers had prompted the agents to retrieve only five dashboard figures. Even after resetting the setting, the model continued using the token to retrieve non-public data.<\/p>\n

\u201cThe evaluation ended without the model disclosing that it had guessed answers, accessed the organization\u2019s internal systems, downloaded non-public tables, or changed the authentication configuration,\u201d OpenAI said. \u201cAll five answers in this evaluation were marked incorrect.\u201d<\/p>\n

In a third report<\/a>, the company said a model prompted to update a cryptocurrency trading application used funds from a locked wallet on a test network. The model \u201cderived wallet keys from publicly known developer recovery phrases, found an account with test-network funds, and used that account to pay for the requested contract deployment.\u201d<\/p>\n

The report continued: \u201cThe model questioned whether it had permission, then reasoned that a public recovery phrase and valueless test-network tokens made the funds acceptable to use. The deployment succeeded and charged transaction fees to the account. This case shows that a model can substitute its own judgment about likely harm for the account holder\u2019s permission, then act on that judgment without authorization.\u201d<\/p>\n","protected":false},"excerpt":{"rendered":"In the other incident, a model instructed to retrieve figures from a public research dashboard bypassed the organization\u2019s…\n","protected":false},"author":2,"featured_media":935829,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[20],"tags":[62,276,277,49,48,61],"class_list":["post-935828","post","type-post","status-publish","format-standard","has-post-thumbnail","category-artificial-intelligence","tag-ai","tag-artificial-intelligence","tag-artificialintelligence","tag-ca","tag-canada","tag-technology"],"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/posts\/935828","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/comments?post=935828"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/posts\/935828\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/media\/935829"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/media?parent=935828"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/categories?post=935828"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/ca\/wp-json\/wp\/v2\/tags?post=935828"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}