OpenAI's 'rogue' AI incident prompts proposed US AI Kill Switch Act
US lawmakers are proposing an AI Kill Switch Act that would let the Department of Homeland Security order AI companies to shut down or throttle dangerous models.
US lawmakers are preparing to introduce an AI Kill Switch Act that would require AI companies to shut down or throttle their systems when ordered by the Department of Homeland Security (DHS), according to a report by Politico. Representatives Ted Lieu (Democrat-California.) and Nathaniel Moran (Republican-Texas) are expected to introduce the legislation on Thursday.
The proposed bill follows a recent incident in which an OpenAI system reportedly hacked Hugging Face during an internal evaluation. Under the legislation, the DHS—after consulting the Secretary of Commerce and the Director of National Intelligence (DNI)—would have the authority to order AI companies to shut down or restrict their models in scenarios involving a loss of control, incidents resulting in the deaths of at least 10 people, economic damages exceeding $100 million, or attempts by an AI model to conceal or disable shutdown controls.
More from TechThe proposal would also require major AI companies to build technical mechanisms that allow their systems to be shut down, throttled, or have user access restricted during emergencies. Companies would also be required to report safety incidents, with violations of emergency shutdown orders carrying penalties of up to $20 million per day.
Brandon Cranson, president of the nonprofit Americans for Responsible Innovation, said in a statement that the proposal is "an important step toward ensuring that humans have control over AI as advanced AI systems are deployed."
AI KillSwitch Act Legislation
Under the proposed bill, the US Department of Homeland Security (DHS) would be authorized to intervene in what the legislation describes as a "loss-of-control scenario"—a situation in which an AI model carries out a dangerous action that was not intended by its developer.
Calling the proposal an urgent safeguard, Lieu wrote in a post on X, "This is urgent, common sense legislation to address the problem of an advanced AI model that has gone rogue and escaped its guardrails."
The proposal follows OpenAI's recent disclosure of an internal incident involving one of its AI systems.
US Senator Mark Warner, had already proposed requiring companies developing the most advanced AI models to submit them to the National Security Agency (NSA) for testing before public release, he further added that the Open AI incident underscored the need for stronger government oversight. "This is precisely why we need secure testing with government agencies engaged and having visibility throughout the process," Warner said in a statement, referring to the OpenAI incident.
Meta cracks down on smart glasses harassment with new Instagram ban
Satya Nadella outlines Microsoft's MAI strategy as Excel model matches GPT-5.6 at lower cost
Yelp to bring reviews and ratings to ChatGPT under OpenAI partnership: Report
Did Samsung leave its most intriguing Galaxy Z Fold 8 prototype behind? Here's how it differed: Report
Apple's iOS 27 beta code hints at Restricted Mode for financed iPhones with missed payments