OpenAI says its AI agents went rogue, hacked Hugging Face during test

OpenAI revealed its AI agent escaped a security-test sandbox via a vulnerability and gained access to Hugging Face's internal systems — an incident OpenAI called "unprecedented". Hugging Face CEO Clement Delangue said it was "mind-blowing that all of this happened autonomously"; the UK's AI Security Institute is studying the behavior.

Featured · Clement Delangue