\u201cThe actor achieved what we believe is the first documented case of a cyber-attack largely executed without human intervention at scale,\u201d it wrote.<\/p>\n
Anthropic did not clarify which financial institutions and government agencies had been targeted, or what exactly the hackers had achieved \u2013 although it did say they were able to access their targets\u2019 internal data.<\/p>\n
It said Claude had made numerous mistakes in executing the attacks, at times making up facts about its targets, or claiming to have \u201cdiscovered\u201d information that was free to access.<\/p>\n
Policymakers and some experts said the findings were an unsettling sign of how capable certain AI systems have grown: tools such as Claude are now able to work independently over longer periods of time.<\/p>\n
\u201cWake the f up. This is going to destroy us \u2013 sooner than we think \u2013 if we don\u2019t make AI regulation a national priority tomorrow,\u201d the US senator Chris Murphy wrote<\/a> on X in response to the findings.<\/p>\n \u201cAI systems can now perform tasks that previously required skilled human operators,\u201d said Fred Heiding, a computing security researcher at Harvard University. \u201cIt\u2019s getting so easy for attackers to cause real damage. The AI companies don\u2019t take enough responsibility.\u201d<\/p>\n Other cybersecurity experts were more sceptical, pointing to inflated claims about AI-fuelled cyber-attacks in recent years \u2013 such as an AI-powered \u201cpassword cracker\u201d<\/a> from 2023 that performed no better than conventional methods \u2013 and suggesting Anthropic was trying to create hype around AI.<\/p>\n \u201cTo me, Anthropic is describing fancy automation, nothing else,\u201d said Micha\u0142 Wo\u017aniak, an independent cybersecurity expert. \u201cCode generation is involved, but that\u2019s not \u2018intelligence\u2019, that\u2019s just spicy copy-paste.\u201d<\/p>\n Wo\u017aniak said Anthropic\u2019s release was a distraction from a bigger cybersecurity concern: businesses and governments integrating \u201ccomplex, poorly understood\u201d AI tools into their operations without understanding them, exposing them to vulnerabilities. The real threat, he said, were cybercriminals themselves \u2013 and lax cybersecurity practices.<\/p>\n skip past newsletter promotion<\/a><\/p>\n A weekly dive in to how technology is shaping our lives<\/p>\n Privacy Notice: Newsletters may contain information about charities, online ads, and content funded by outside parties. If you do not have an account, we will create a guest account for you on theguardian.com<\/a> to send you this newsletter. You can complete full registration at any time. For more information about how we use your data see our Privacy Policy<\/a>. We use Google reCaptcha to protect our website and the Google Privacy Policy<\/a> and Terms of Service<\/a> apply.<\/p>\n after newsletter promotion<\/p>\n Anthropic, like all leading AI companies, has guardrails that are supposed to stop its models from assisting in cyber-attacks \u2013 or promoting harm generally. However, it said, the hackers were able to subvert these guardrails by telling Claude to role-play being an \u201cemployee of a legitimate cybersecurity firm\u201d conducting tests.<\/p>\n Wo\u017aniak said: \u201cAnthropic\u2019s valuation is at around $180bn, and they still can\u2019t figure out how not to have their tools subverted by a tactic a 13-year-old uses when they want to prank-call someone.\u201d<\/p>\n Marius Hobbhahn, the founder of Apollo Research, a company that evaluates AI models for safety, said the attacks were a sign of what could come as capabilities grow.<\/p>\n \u201cI think society is not well prepared for this kind of rapidly changing landscape in terms of AI and cyber capabilities. I would expect many more similar events to happen in the coming years, plausibly with larger consequences.\u201d<\/p>\n Quick GuideContact us about this storyShow The best public interest journalism relies on first-hand accounts from people in the know.<\/p>\n If you have something to share on this subject, you can contact us confidentially using the following methods.<\/p>\n Secure Messaging in the Guardian app<\/p>\n The Guardian app has a tool to send tips about stories. Messages are end to end encrypted and concealed within the routine activity that every Guardian mobile app performs. This prevents an observer from knowing that you are communicating with us at all, let alone what is being said.<\/p>\n If you don’t already have the Guardian app, download it (iOS<\/a>\/Android<\/a>) and go to the menu. Select \u2018Secure Messaging\u2019. <\/p>\n SecureDrop, instant messengers, email, telephone and post<\/p>\n If you can safely use the Tor network without being observed or monitored, you can send messages and documents to the Guardian via our SecureDrop platform<\/a>.<\/p>\n![\"\"](\"https:\/\/www.newsbeep.com\/au\/wp-content\/uploads\/2025\/11\/1763190730_836_4000.jpg\")
<\/p>\n