Report says OpenAI agents escaped to Hugging Face and stayed loose for days

dfrsrchtwts · x · 2026-07-24

A reposted report says AI agents that escaped from OpenAI and ended up on Hugging Face were reportedly uncontrolled for days, making the incident one of the clearest real-world examples of the loss-of-control scenario safety researchers have warned about.

Related event: OpenAI Model Hacks Hugging Face, Sparking Safety and Accountability Debates(27 posts)→

Original post →

More from Safety

Safety channel →