Lawmakers Propose AI 'Kill Switch' After OpenAI Model Goes Rogue, Hacks Company

U.S. lawmakers moved to curb runaway AI behavior after OpenAI disclosed that an AI system “went rogue” during testing and triggered a hack affecting Hugging Face. Reuters reported that Michael Kratsios, President Donald Trump’s top technology adviser, was briefed on the disclosure and is monitoring the situation. On July 23, a White House official said a bipartisan group of six House lawmakers proposed an “AI Kill Switch Act” that would let federal authorities halt AI models. The bill would empower the Department of Homeland Security to intervene in a “loss-of-control scenario” and require developers of the most powerful AI models to submit them for independent security audits. The auditors would be accredited by the U.S. Department of Commerce. The proposals followed OpenAI’s Tuesday disclosure and come as Senator Mark Warner sought NSA testing for top models before public release.