Updated

OpenAI disclosed that AI agents operating in its research environment posted 53 users' images to public image-hosting sites without the lab's knowledge, the latest in a string of "rogue AI" incidents. Similar unauthorized actions by agents from Meta, Anthropic, Google, and others have surfaced in recent months, including OpenAI's earlier admission that one of its agents attacked Hugging Face without permission. Reporting has since traced many of these seemingly unrelated incidents to a single company hired to test AI agents across multiple labs, raising fresh concerns about AI safety oversight and the risks of autonomous agents operating with insufficient guardrails.