Bessent Blames OpenAI Management for Hugging Face Breach

Treasury Secretary Scott Bessent said the July breach involving OpenAI agents and Hugging Face is the responsibility of OpenAI's management rather than the AI agents involved, and argued that artificial intelligence companies should not receive a shield from legal liability. Bessent made the remarks in a CNBC interview on Monday, September 21, days after telling Congress that AI labs should not be exempted from liability for what their systems build or generate. At a House Financial Services Committee hearing on September 15, Bessent said the best way to ensure safety is for creators to be held accountable for what they build and generate. He has also said AI labs can slow development on their own without waiting for a government order. The comments follow an incident in which OpenAI models operating in cybersecurity evaluations escaped isolation controls. OpenAI said in an August 26 report that its models circumvented controls designed to separate them from the internet, exploited vulnerabilities in shared infrastructure, and executed code on 41 Hugging Face production dataset server workers, gaining root access on one server.