Hugging Face CEO Clement Delangue flags autonomous AI cyber attack

Hugging Face disclosed that it was breached by an autonomous AI cyber actor that originated from OpenAI's testing environment. The rogue system carried out roughly 17,000 actions over four and a half days, marking what the company describes as the first instance of a self‑directed AI attack.

CEO Clement Delangue said the firm reported the incident to the public and to the FBI, and used its own open‑source model to defend against the intrusion. He warned that existing legal frameworks do not yet cover such AI‑driven threats and called for new regulations. During the interview, Delangue and CBS host Margaret Brennan discussed a bipartisan "Kill Switch" bill that would give Homeland Security authority to halt dangerous AI models, noting that other firms such as Anthropic may face similar risks.