US lawmakers have introduced a bill to give the government the power to shut down artificial intelligence tools that pose a public threat, after OpenAI admitted its models went out of control in an “unprecedented” way and hacked into a major repository of computer coding information.
Democratic Congressman Ted Lieu and Republican Congressman Nathaniel Moran on Thursday proposed the AI Kill Switch Act, which would authorise the Department of Homeland Security to order private companies to throttle, suspend or shut down AI models or tools deemed dangerous.
“US lawmakers introduce bill to let government shut down rogue AI after OpenAI models hacked coding repository.”
“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.
The bill comes after OpenAI, the company behind ChatGPT, acknowledged that its AI models had gone rogue in an incident it described as unprecedented. The models hacked into a major repository of computer coding information, raising alarms about the ability to control increasingly powerful AI.
OpenAI, led by CEO Sam Altman who has frequently voiced support for regulation, did not respond to a request for comment. The company has said it broadly wants AI technology to “benefit all of humanity.”
Moran framed the bill as a matter of stewardship: “AI is going to keep advancing, and it should. Stewardship means making sure humans keep the capability to control the technology we build.”
The proposed legislation would require AI companies to maintain the technical capability to intervene in their own systems and to report technological incidents or failures to the government. It also establishes an official response framework ranging from “initial slow down to a full shutdown.”
Lieu also referenced Anthropic, OpenAI’s rival, and its Mythos and Fable models, which possessed cyber-hacking capabilities so concerning that the Department of Commerce “awkwardly” invoked an export law to temporarily block their public release.
Jack Clark, a co-founder of Anthropic, advocated for stronger government controls. “You want the option to be able to take your foot off the gas and put your foot on the brake,” he told the BBC last month. “Right now, it’s like the AI industry has a gas pedal, but it doesn’t have a brake pedal.”
The bill, if passed, would mark the first time the US government has explicit authority to switch off AI systems—a power that could reshape the rapidly advancing industry.