Sign up to see the future, today<\/p>\n
Can\u2019t-miss innovations from the bleeding edge of science and tech<\/p>\n
A rogue AI agent caused a critical security incident at Meta which exposed sensitive users data to people who didn\u2019t have proper authorization, according to reporting from The Information<\/a> and The Verge<\/a>, in the latest illustration of the safety pitfalls endemic to AI systems. <\/p>\n The blunder occurred last week when a software engineer used an in-house AI agent to break down a technical question posed by another employee on an internal discussion forum, per company communications and an incident report. The in-house AI was likened to OpenClaw<\/a>, an open source agentic model that\u2019s generated loads of hype in tech circles for being an AI that\u00a0\u201cactually does things.\u201d<\/p>\n What transpired was a mix of AI hallucination and a game of telephone<\/a>. The AI posted its response to the forum without the approval of the employee who prompted it. Then another employee acted on the AI\u2019s advice, which turned out to contain \u201cinaccurate information.\u201d<\/p>\n A mini crisis unfolded. For almost two hours, unauthorized access to troves of sensitive company and user data was given to engineers who weren\u2019t approved to view the data before. Meta classified the screw-up as a \u201cSEV1\u201d level incident, the second highest level of severity on a scale the company uses to rank security incidents.<\/p>\n For now, it doesn\u2019t appear anything nefarious happened as a result of the unauthorized access, and a Meta spokesperson told The Verge that \u201cno user data was mishandled.\u201d<\/p>\n The spokesperson emphasized that the AI agent itself didn\u2019t make any technical changes, shifting the blame to human error.<\/p>\n \u201cThe employee interacting with the system was fully aware that they were communicating with an automated bot. This was indicated by a disclaimer noted in the footer and by the employee\u2019s own reply on that thread,\u201d they told The Verge. \u201cThe agent took no action aside from providing a response to a question. Had the engineer that acted on that known better, or did other checks, this would have been avoided.\u201d<\/p>\n Whether Meta is incentivized to downplay the incident due to embarrassment or play it up to build hype about AI\u2019s emerging capabilities is anybody\u2019s guess.<\/p>\n But at its tech rivals, AI agents have been responsible for making catastrophic technical changes. At least two outages at Amazon Web Services last year were caused when Amazon\u2019s in-house AI coding tool made erroneous changes<\/a>, including deleting the entire coding environment. Amazon leaders admitted<\/a> in a March meeting that \u201cgen-AI assisted changes\u201d were disrupting its core e-commerce business, and insisted that going forward there would be more oversight on how AI-coding changes are implemented.<\/p>\n Perhaps presaging this latest incident was an AI-related mistake that a senior Meta employee admitted to last month. In a widely mocked post<\/a>, its director of AI safety Summer Yue said that an OpenClaw agent she was experimenting with \u2014 by giving it control of her personal computer \u2014 nearly wiped out her entire email inbox<\/a> while ignoring her instructions to stop.<\/p>\n