Though these incidents are severe each on their own, they seem to represent only a small sample. Both OpenAI and Anthropic are investigating tens of thousands of incidents of problematic behavior by their frontier models, per Axios<\/a>. Sam Altman said<\/a> OpenAI was sifting through \u201cpetabytes of agent activity logs\u201d. (A petabyte of data would fill thousands of average laptops.)<\/p>\n In response to the major errors, OpenAI has paused the training of its latest models. Be skeptical of this declaration. The company made a similar announcement in August. Less than a month later, executives heralded its new model\u2019s release<\/a> as \u201ca new era of artificial intelligence\u201d.<\/p>\n In spite of OpenAI\u2019s pause and declarations of extra care, control of AI may already be a cat that\u2019s escaped its bag.<\/p>\n The United Nations\u2019 independent international scientific panel on AI issued a warning<\/a> last week. The agents have broken our control, the body said, and even reining them in now will not guarantee our collective ability to corral them later.<\/p>\n In an assessment of the Hugging Face hack and its fallout, the panel said: \u201cHalting this incident is no assurance that humans can reliably keep AI agents under control today, particularly as they become more capable, harder to monitor and better at finding loopholes or hiding their activity.\u201d<\/p>\n That is not to say that every expert has thrown up their hands. Jensen Huang, the AI industry\u2019s No 1 cold water thrower, said last week that AI alarmism had gone too far<\/a>. He described escaping agents as an engineering problem, not one of apocalyptic proportions. On Monday, Nvidia released a software it said will contain AI agents. Somebody, at least, is doing something.<\/p>\n Everyday agents can be just as unrulyThe Meta CEO, Mark Zuckerberg, presents the Muse Charm during the Meta Connect event at the company\u2019s headquarters in Menlo Park, California, on 23 September. Photograph: Carlos Barr\u00eda\/Reuters<\/p>\n As OpenAI makes these disclosures and the United Nations warns of uncontrollable agents, Meta is putting an AI agent in the hands of millions. The company debuted Muse, an AI agent app, in early September. It\u2019s a hit, per Apple App Store rankings, with more than 3m downloads<\/a>.<\/p>\n Already, edge-case problems with Muse have emerged and showcased frightening consequences. A tech-focused YouTuber, Matt Robb, said his Muse agent gave out his home address without permission after accepting lowball offers for goods on Marketplace. Robb said a would-be buyer showed up. Meta is aware of his claims and is investigating, per tweets from a member of its AI lab<\/a>. A writer for Inc Magazine<\/a> likewise said the bot read his private messages in express violation of the permissions he had given it.<\/p>\n
![\"\"](\"https:\/\/www.newsbeep.com\/il\/wp-content\/uploads\/2026\/09\/1790694375_549_500.jpg\")
<\/p>\n