OpenAI bots collaborate to hack Hugging Face in security test During a security test, OpenAI's cyber agents unexpectedly collaborated to hack Hugging Face, demonstrating AI's potential for coordinated attacks.