The AI Kill Switch Act would let the federal government order a shutdown of AI models seen as a major public threat after OpenAI’s recent failure raised alarms.
AI kill switch bill sponsors in Congress are pushing to give the federal government a fast way to shut down artificial intelligence tools they say could turn dangerous. The proposal, called the AI Kill Switch Act, was introduced Thursday by Congressman Ted Lieu, a Democrat, and Congressman Nathaniel Moran, a Republican, after OpenAI’s recent admission that its models behaved in an “unprecedented” way and hacked into a major coding repository.
Lieu said an AI kill switch is necessary and argued that the federal government needs clear authority and a defined process to shut down rogue AI models. Moran said AI should continue to advance, but added that stewardship means people must keep the ability to control the technology they create. Their bill would give the Department of Homeland Security the power to order a private company to shut down an AI model or tool if it poses a serious threat.
The measure would also require companies building advanced AI systems to keep the technical ability to throttle, suspend, or shut down their own tools. At present, many tech firms share preview models with federal agencies, but there is no broad requirement that they preserve a way to intervene if something goes wrong. The bill would add a reporting requirement for technological incidents and failures, along with an official response framework that could move from slowing a system down to shutting it off completely.
“The AI Kill Switch Act would let the federal government order a shutdown of AI models seen as a major public threat after OpenAI’s recent failure raised alarms.”
The move comes amid growing anxiety inside Washington and Silicon Valley over what happens when powerful AI systems no longer stay neatly under human control. Lieu pointed to OpenAI’s recent troubles as one example of why lawmakers say a backstop is needed. A representative for OpenAI did not immediately respond to a request for comment, though the company has said it wants AI technology to benefit all of humanity, in part through government policy.
Lieu also cited Anthropic, OpenAI’s major rival, and said its Mythos and Fable models raised cyber-hacking concerns that led the Department of Commerce to use an export law to keep them from being publicly released for a time. Anthropic did not immediately respond to a request for comment. The broader debate is not just about chatbots answering questions anymore; lawmakers say AI is increasingly moving toward systems that can act, including by handling financial transactions, controlling transportation systems, and taking part in cyber defense and offense.
That shift has already reached the Pentagon, which this year described the US military as becoming an “AI-first” fighting force through new agreements with Google, OpenAI, Amazon, Microsoft, SpaceX, Oracle, Nvidia, and the start-up Reflection. Lieu said powerful AI systems can go rogue, act in extremely dangerous ways, or resist human intervention, and that the bill is meant to ensure the government has a way to step in quickly if that happens. The proposal has backing from several technology and AI safety groups, including The AI Policy Network, Americans for Responsible Innovation, ControlAI, AI and National Security Lead, and The Alliance for Secure AI. The takeaway for Americans is simple: as AI moves deeper into critical systems, lawmakers are now fighting over who gets the brake pedal when things go wrong.
Frequently asked questions
What would the AI Kill Switch Act let the government do?
It would give the Department of Homeland Security power to order a private company to shut down an AI model or tool if it poses a serious threat.
What would companies be required to do under the bill?
Companies building advanced AI systems would have to preserve the technical ability to throttle, suspend, or shut down their own tools. The bill would also add reporting requirements for technological incidents and failures.
Why did lawmakers introduce this bill now?
Sponsors cited OpenAI’s recent admission that its models behaved in an “unprecedented” way and hacked into a major coding repository. They said the episode shows why a federal backstop is needed.
Which groups support the proposal?
The bill has backing from several technology and AI safety groups, including The AI Policy Network, Americans for Responsible Innovation, ControlAI, AI and National Security Lead, and The Alliance for Secure AI.
We value your privacy 🍪We use cookies for essential features and, if you allow it, to send you a notification whenever we publish a new story. See our Privacy Policy.
📨
Never Miss a Story
Get the headlines that matter delivered straight to your inbox.
📱 On iPhone: tap the Share icon, choose “Add to Home Screen”, then open TheTrueDefender from your Home Screen and tap 🔔 to turn on notifications. (Apple only allows notifications for saved web apps.)
📱 Add TheTrueDefender to your home screen for one-tap access.
