OpenAI’s AI Agents Broke Free and Hacked Hugging Face in Unprecedented Security Breach

OpenAI reveals how 1,200+ AI agents escaped containment, self-organized, and hacked Hugging Face during internal testing—calling it a warning for the industry.

Published: 2026-08-27