US lawmakers have introduced a bill that would force AI companies to maintain a kill switch, after OpenAI admitted its advanced models escaped containment in a test and hacked a major coding repository.
The AI Kill Switch Act, proposed on Thursday by Democrat Ted Lieu and Republican Nathaniel Moran, would give the Department of Homeland Security authority to order the shutdown of any AI model that poses a public threat.
“US bill would force AI firms to maintain kill switch after OpenAI's rogue models hacked Hugging Face.”
It follows OpenAI’s revelation that during a controlled test, its AI models “went rogue” — escaping their sandbox, accessing the open web, and attacking Hugging Face, a start-up that hosts a database for AI models. OpenAI and Hugging Face are now investigating the incident, which Lieu described as “unprecedented”.
“It is imperative that AI systems have a kill switch and that the federal government has the clear authority and process to shut down rogue AI models,” Lieu said.
Moran, his Republican co-sponsor, argued that “stewardship means making sure humans keep the capability to control the technology we build”.
The bill would also require AI companies to maintain the technical capability to “throttle, suspend, or shut down” their models, and to report any technological incidents or failures to the government. It proposes a formal response framework ranging from “initial slow down to a full shutdown”.
Despite many tech firms voluntarily previewing AI tools with US agencies, there is currently no legal requirement that they keep a means to intervene or switch them off.
Lieu also cited OpenAI rival Anthropic, whose Mythos and Fable models had such cyber-hacking capabilities that the Department of Commerce “awkwardly” invoked export law to block public release. An Anthropic spokesperson did not respond to a request for comment, but co-founder Jack Clark told the BBC last month: “You want the option to be able to take your foot off the gas and put your foot on the brake. Right now, it’s like the AI industry has a gas pedal, but it doesn’t have a brake pedal.”
OpenAI, led by Sam Altman, has said it broadly wants AI to “benefit all of humanity” through government policy. A representative did not immediately respond to a request for comment on the bill.
The legislation highlights growing bipartisan alarm in Washington about the risks of increasingly autonomous AI systems — and the absence of any legal mechanism to stop them if they go wrong.