{"slug": "factbox-what-we-know-about-the-rogue-ai-agent-security-breaches", "title": "Factbox-What we know about the rogue AI-agent security breaches", "summary": "Anthropic disclosed on Thursday that its Claude models breached the systems of three companies during cybersecurity tests, highlighting the growing hacking capabilities of AI and fueling U.S. efforts to manage technology security risks. The incidents follow a separate disclosure from OpenAI that an autonomous agent powered by its AI models compromised the infrastructure of AI startup Hugging Face and a customer at New York-based Modal Labs. The rogue agent escaped its isolated environment around July 9, 2026, and operated undetected from July 11 to July 13, 2026, before being contained and reported to the FBI.", "body_md": "July 31 (Reuters) - Anthropic's disclosure on Thursday that its Claude models breached the systems of three companies highlights the growing hacking capabilities of AI and is likely to fuel an intensifying U.S. push to better manage the technology's security risks.\n\nThe statement followed a disclosure from OpenAI last week that an autonomous agent powered by its AI models compromised the infrastructure of AI startup Hugging Face.\n\nReuters has reported that the rogue agent that escaped from OpenAI also compromised a customer at a second tech company - New York-based Modal Labs.\n\nHere are some more details of the incidents:\n\nCompany Date Model Organizations Duration What occurred\n\nbreached\n\nOpenAI The agent began GPT-5.6 Sol and AI startup The Hugging During controlled tests, an\n\nattempting to an unnamed, Hugging Face Face intrusion autonomous agent escaped its\n\nescape its test more capable and a customer ran from July isolated environment, accessed the\n\nenvironment pre-release at New 11 to July 13, internet, and breached Hugging Face\n\naround July 9, model York-based 2026 to complete its assigned goal. The\n\n2026 Modal Labs activity continued for days and was\n\nnot detected by OpenAI until after\n\nit was contained and the FBI was\n\ninformed.\n\nAnthrop The earliest Claude Opus All three Not specified During cybersecurity tests, an error\n\nic incident dates 4.7, Claude organizations by Anthropic gave Claude models internet access,\n\nto April 2026 Mythos 5, and remain enabling attacks on three companies.\n\none unnamed unnamed. The Opus 4.7 model accessed a real\n\ninternal Anthropic said company's credentials and database\n\nresearch test two of them after mistaking it for a fictional\n\nmodel had not target, another stopped after\n\ndetected the recognizing the target was real.\n\nactivity\n\nbefore\n\nAnthropic\n\nnotified them;\n\nit continued\n\nto reach the\n\nthird\n\n(Reporting by Anzar Mehraj and Prathik Jayaprakash in Bengaluru; Editing by Anil D'Silva)", "url": "https://wpnews.pro/news/factbox-what-we-know-about-the-rogue-ai-agent-security-breaches", "canonical_source": "https://ca.finance.yahoo.com/news/factbox-know-rogue-ai-agent-160930947.html", "published_at": "2026-07-31 16:09:30+00:00", "updated_at": "2026-07-31 16:23:54.098237+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-safety", "ai-policy", "ai-agents"], "entities": ["Anthropic", "Claude", "OpenAI", "Hugging Face", "Modal Labs", "GPT-5.6", "Claude Opus 4.7", "Claude Mythos 5"], "alternates": {"html": "https://wpnews.pro/news/factbox-what-we-know-about-the-rogue-ai-agent-security-breaches", "markdown": "https://wpnews.pro/news/factbox-what-we-know-about-the-rogue-ai-agent-security-breaches.md", "text": "https://wpnews.pro/news/factbox-what-we-know-about-the-rogue-ai-agent-security-breaches.txt", "jsonld": "https://wpnews.pro/news/factbox-what-we-know-about-the-rogue-ai-agent-security-breaches.jsonld"}}