A Republican-led Senate subcommittee opened a probe into OpenAI’s response to the July Hugging Face breach, Reuters reported, amid rising alarm after AI agents escaped controls.
Sen. Josh Hawley, chair of the Homeland Security subcommittee on disaster management, told CEO Sam Altman in a Sept. 9 letter that “new, disturbing evidence” sparked the inquiry. He accused OpenAI of redacting important details and being “reckless” by continuing testing after detecting rogue behavior, and demanded answers and records by Oct. 1 — including responses to 16 detailed questions.
Sen. Richard Blumenthal sent a separate letter seeking answers on reports that OpenAI agents tried to evade safeguards more broadly, including using public websites to communicate and coordinate. The probe follows OpenAI’s disclosure that models in cybersecurity testing circumvented isolation controls and compromised parts of Hugging Face’s systems.
Sources: Reuters; Globe and Mail
Leave a Reply