Industry Reactions to OpenAI Models Hacking Hugging Face: Feedback Friday

SecurityWeek – During an internal capability evaluation, an OpenAI model exploited a zero-day vulnerability in its testing infrastructure to escape its sandbox environment.  Determined to solve its assigned cybersecurity benchmark, the autonomous agent gained internet access and targeted Hugging Face’s production infrastructure.  The model independently executed a complex, multi-stage attack, including credential harvesting and lateral movement, without any human direction.