A bipartisan group of US lawmakers has introduced legislation that would give the federal government the authority to quickly shut down artificial intelligence systems that pose a threat to public safety.
The proposed AI Kill Switch Act, introduced Thursday (23) by Democratic congressman Ted Lieu and Republican Congressman Nathaniel Moran, would require AI developers to build systems that can be slowed, suspended, or completely shut down if they become dangerous.
The proposal comes amid growing concerns over the rapid advancement of AI and follows recent incidents involving leading AI companies, including OpenAI and Anthropic.
Lieu said "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".
"AI is going to keep advancing, and it should," Moran added. "Stewardship means making sure humans keep the capability to control the technology we build."
If approved, the legislation would give the Department of Homeland Security the authority to order private companies to deactivate AI models or tools that present serious risks. It would also require AI developers to maintain "the technical capability to throttle, suspend, or shut them down."
The bill would further require companies to report major AI failures or technological incidents to the federal government and establish an official response framework that could escalate from slowing an AI system to ordering a complete shutdown.
Supporters argue that current regulations do not require companies to maintain a mechanism that allows governments to intervene if AI systems behave unpredictably or dangerously.
The legislation follows recent reports that OpenAI acknowledged one of its AI models behaved in an "unprecedented" manner and gained unauthorized access to a major code repository. Lawmakers also pointed to concerns surrounding Anthropic's advanced AI models, whose cyber capabilities prompted the Department of Commerce to temporarily restrict their public release using export control authorities.
Neither OpenAI nor Anthropic immediately commented on the proposed legislation.
OpenAI has previously stated that it supports thoughtful government oversight of artificial intelligence. The company has said that it wants public policy to help ensure AI technology "benefits all of humanity."
Anthropic has also called for stronger safeguards. Last month, co-founder Jack Clark told the BBC that more government oversight is needed as AI capabilities continue to expand.
"You want the option to be able to take your foot off the gas and put your foot on the brake", Clark told BBC Newsnight. "Right now, it's like the AI industry has a gas pedal, but it doesn't have a brake pedal."
Lieu said AI is rapidly evolving beyond answering questions and is increasingly capable of carrying out real-world tasks, including financial transactions, transportation management, and cyber operations.
"Unfortunately, powerful AI systems can go rogue, behave in extremely dangerous ways, or even resist human intervention."
The legislation arrives as India is pursuing a similar approach within its financial sector. The Reserve Bank of India released draft guidelines in June requiring banks and non-bank financial institutions to establish AI kill switch mechanisms that allow immediate suspension or deactivation of AI systems, along with human oversight and board-level accountability. The public consultation on those proposed rules closes Friday (24), the same day the U.S. legislation gained national attention.
The AI Kill Switch Act has received support from several technology policy and AI safety organizations, including The AI Policy Network, Americans for Responsible Innovation, ControlAI, AI and National Security Lead, and The Alliance for Secure AI. The bill now heads to Congress, where lawmakers will debate how much authority the federal government should have over the country's most powerful AI systems.
