{"id":292920,"date":"2025-11-15T09:26:33","date_gmt":"2025-11-15T09:26:33","guid":{"rendered":"https:\/\/www.newsbeep.com\/us\/292920\/"},"modified":"2025-11-15T09:26:33","modified_gmt":"2025-11-15T09:26:33","slug":"anthropic-says-it-disrupted-what-it-calls-the-first-documented-case-of-a-large-scale-ai-cyberattack-executed-without-substantial-human-intervention","status":"publish","type":"post","link":"https:\/\/www.newsbeep.com\/us\/292920\/","title":{"rendered":"Anthropic says it ‘disrupted’ what it calls ‘the first documented case of a large-scale AI cyberattack executed without substantial human intervention’"},"content":{"rendered":"

Anthropic, the $183 billion San Francisco\u2013based AI company<\/a> known for the Claude chatbot, said it thwarted what it called the first documented, large-scale cyberattack orchestrated predominantly by artificial intelligence. The attack, it said on X<\/a>, \u201chas significant implications for cybersecurity in the age of AI agents.\u201d<\/p>\n

Anthropic released a blog post about the incident<\/a> on Thursday. The company said it detected \u201csuspicious activity\u201d in mid-September that, upon investigation, showed \u201ca highly sophisticated espionage campaign.\u201d <\/p>\n

\u201cAccording to the company, \u201c\u201dThe attackers used AI\u2019s \u2018agentic\u2019 capabilities to an unprecedented degree\u2014using AI not just as an advisor, but to execute the cyberattacks themselves,\u201d the company said. <\/p>\n

Anthropic said, with \u201chigh confidence,\u201d it identified the threat actor as a Chinese state-sponsored group that successfully manipulated its Claude Code tool into attempting to infiltrate about 30 global targets, including large tech companies, financial institutions, chemical manufacturers, and government agencies.<\/p>\n

The attackers, Anthropic said, \u201cbroke down their attacks into small, seemingly innocent tasks that Claude would execute without being provided the full context of their malicious purpose.\u201d <\/p>\n

To bypass the system\u2019s safeguards, the attackers allegedly posed as a legitimate cybersecurity firm conducting defensive testing and successfully \u201cjailbroke\u201d Claude, enabling it to operate beyond its safety guardrails. This allowed the AI not just to assist, but to autonomously inspect digital infrastructure, identify \u201cthe highest-value databases,\u201d write exploit code, harvest user credentials, and organize stolen data\u2014\u201call with minimal human supervision,\u201d according to Anthropic.\u200b<\/p>\n

In response, the company said it immediately began mapping the scope of the operation, banned the attackers\u2019 accounts as they were identified, notified affected organizations, and coordinated with authorities over a ten-day investigation. <\/p>\n

Anthropic said it has also upgraded its detection systems, developing classifiers to flag and prevent similar attacks, and has committed to sharing such case studies publicly \u201cto help those in industry, government, and the wider research community strengthen their own cyber defenses.\u201d\u200b<\/p>\n

Most notably, the company said the vast majority\u2014roughly \u201c80-90%\u201d\u2014of the work done in this particular cyberattack was executed by AI. <\/p>\n

\u201cThe sheer amount of work performed by the AI would have taken vast amounts of time for a human team. At the peak of its attack, the AI made thousands of requests, often multiple per second\u2014an attack speed that would have been, for human hackers, simply impossible to match,\u201d the company said.\u200b<\/p>\n

Anthropic did mention that a fully autonomous cyberattack is still likely a pipe dream, at least for now, as Claude occasionally \u201challucinated credentials or claimed to have extracted secret information that was in fact publicly available.\u201d But the company made it clear\u200b \u201cthe barriers to performing sophisticated cyberattacks have dropped substantially\u2014and we predict that they\u2019ll continue to do so.\u201d<\/p>\n

\u201cWith the correct setup, threat actors can now use agentic AI systems for extended periods to do the work of entire teams of experienced hackers: analyzing target systems, producing exploit code, and scanning vast datasets of stolen information more efficiently than any human operator,\u201d it wrote. \u201cLess experienced and resourced groups can now potentially perform large-scale attacks of this nature.\u201d\u200b<\/p>\n","protected":false},"excerpt":{"rendered":"Anthropic, the $183 billion San Francisco\u2013based AI company known for the Claude chatbot, said it thwarted what it…\n","protected":false},"author":2,"featured_media":292921,"comment_status":"","ping_status":"","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[45],"tags":[182,184,181,507,144,5477,7270,74],"class_list":{"0":"post-292920","1":"post","2":"type-post","3":"status-publish","4":"format-standard","5":"has-post-thumbnail","7":"category-artificial-intelligence","8":"tag-ai","9":"tag-anthropic","10":"tag-artificial-intelligence","11":"tag-artificialintelligence","12":"tag-china","13":"tag-national-security","14":"tag-security","15":"tag-technology"},"_links":{"self":[{"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/posts\/292920","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/comments?post=292920"}],"version-history":[{"count":0,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/posts\/292920\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/media\/292921"}],"wp:attachment":[{"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/media?parent=292920"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/categories?post=292920"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.newsbeep.com\/us\/wp-json\/wp\/v2\/tags?post=292920"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}