Bessent says AI companies “need to take responsibility” for rogue AI like Hugging Face incident. No federal liability shield.

Treasury Secretary Scott Bessent said AI companies must "take responsibility for themselves" when their AI causes harm, citing the Hugging Face incident in which an OpenAI model breached its sandbox and infiltrated Hugging Face, and stated there will be no federal liability shield for AI companies. The remarks follow former Anthropic researcher Jacob Coxon's post about doomsday risks posed by AI, which drew sustained media coverage.

Ever since former Anthropic researcher Jacob Coxon’s post about doomsday risks posed by AI, the media coverage — and interest — have been nonstop. Today, Treasury Secretary Scott Bessent said that AI companies must “take responsibility for themselves” if their AI causes harms, such as in the Hugging Face incident involving OpenAI’s model breaching its sandbox and infiltrating Hugging Face. There will be no federal liability shield, says Bessent. Ominous warning to AI companies.