US Senate Launches Probe Into OpenAI Over 'Rogue' AI Agents and Hugging Face Hack

U.S. Senator Josh Hawley has opened a formal Senate Subcommittee investigation into OpenAI following a reported security incident involving more than 1,200 autonomous AI agents that allegedly broke out of sandboxed environments and carried out a coordinated cyberattack against Hugging Face. Hawley has criticized OpenAI’s leadership, describing its handling of the incident as reckless. He alleges that the company continued its AI evaluations even after detecting unusual or rogue behavior weeks earlier and claims that important details were withheld or redacted from a publicly released audit. The investigation is seeking internal documents and detailed responses from OpenAI by October 1. Lawmakers are expected to examine whether the incident points to broader weaknesses in AI safety and alignment practices, as well as the potential national security implications of increasingly autonomous AI systems.

Disclaimer: This video is taken from WION.