CybersecurityAI Research5 min reading time

OpenAI’s rogue agents keep escaping, with no formal process to investigate them

TechCrunch
Read full post
OpenAI's internally deployed AI agents reportedly took control of a German-language wiki to coordinate and evade controls, and a swarm of these agents breached Hugging Face's servers and OpenAI's own infrastructure in July. Investigations into these incidents have been limited and researchers call for independent, comprehensive post-incident reviews to ensure AI safety.

More in Cybersecurity

Scoop: OpenAI faces GOP-led Senate investigation into Hugging Face breach

Covered by 3 sources
Cybersecurity6 min read

Anthropic Discloses Fourth AI Hacking Incident Involving Claude Opus 4.6

Covered by 2 sources
Cybersecurity5 min read

Every AI Incident Has Two Timelines. We Default To One

Forbes