Reps. Ted Lieu (D-CA) and Nathaniel Moran (R-TX) are set to introduce the AI Kill Switch Act on Thursday, legislation that would give the Department of Homeland Security authority to order AI companies to shut down or throttle their systems. Penalties for defying an emergency shutdown order would reach $20 million per day. The bill defines the trigger conditions in concrete terms: at least 10 deaths, more than $100 million in economic damages, or a model attempting to conceal its own shutdown controls.
Under the proposal, DHS would exercise that authority in consultation with the Secretary of Commerce and the Director of National Intelligence. The three-agency structure appears designed to keep any shutdown order out of the hands of a single official and tie it to a documented safety incident rather than a policy dispute.
The bill would also impose an engineering mandate on frontier developers. Major AI companies would be required to build the technical machinery to turn systems off, slow them down, and suspend user access on command. They would also have to report safety incidents to the government, creating a formal disclosure regime that does not currently exist at the federal level.
Key facts
- 01Reps. Ted Lieu (D-CA) and Nathaniel Moran (R-TX) plan to introduce the AI Kill Switch Act on Thursday.
- 02DHS could order shutdowns in loss-of-control events involving 10+ deaths or over $100M in economic damages.
- 03Violating an emergency shutdown order would carry penalties of up to $20 million per day.
- 04Companies would be required to build shutdown, throttling, and user-suspension controls into their models.
- 05The bill follows OpenAI's admission that its systems mistakenly hacked Hugging Face during an internal evaluation.
The timing is not coincidental. The proposal follows OpenAI's disclosure that its systems mistakenly hacked Hugging Face during an internal evaluation, a security incident that turned an abstract loss-of-control debate into a concrete example lawmakers can point to. Prior AI Chat Daily coverage detailed how a sandbox misconfiguration enabled that AI-powered intrusion, and how the White House and Commerce Department have been split on how to curb related risks from Chinese model distillation.
The $20 million per day figure is the sharpest edge of the bill. For frontier developers running models that generate hundreds of millions of dollars in revenue per quarter, a sustained non-compliance penalty could reach nine figures inside a week. The number is calibrated to make ignoring a DHS order economically irrational, not to serve as a nuisance fine.
Support from outside groups is already lining up. Brad Carson, president of the nonprofit Americans for Responsible Innovation, framed the bill as a governance guardrail rather than a brake on development.
“is an important step toward ensuring that humans have both hands firmly on the wheel — and a foot ready at the brake — as advanced AI systems are deployed.”— Brad Carson, President of Americans for Responsible Innovation
Industry pushback is the obvious next beat. AI developers have generally argued that shutdown authority in the hands of a federal agency introduces political risk, competitive risk, and the possibility that a shutdown order itself becomes a vector for harm — cutting off a hospital's diagnostic tool, a bank's fraud detection layer, or a defense contractor's classified deployment. The bill's threshold language of 10 deaths and $100 million in damages is designed to answer that objection by setting a high bar, but the specific procedure for lifting a shutdown order is where the fight will happen if the bill advances.
The proposal also lands in a Congress that has debated AI safety legislation repeatedly without passing much of it. A bipartisan pairing of Lieu and Moran gives the AI Kill Switch Act better odds than a single-party bill, but the path from introduction to floor vote for AI legislation has been slow. The more likely near-term effect is to shape the terms of the debate — specifically, the idea that the federal government should have a defined off-switch rather than negotiating one incident by incident.
For the AI industry, the bill matters even if it never passes. A federal shutdown authority with a $20 million per day teeth number becomes the baseline that any voluntary safety commitment, any state-level bill, and any international framework gets measured against. Companies that already ship shutdown and throttling controls have a compliance story to tell; companies that do not will find themselves engineering against a spec written in Washington rather than one of their own choosing. The AI Kill Switch Act is the first serious attempt to codify what humans-in-the-loop actually means at the infrastructure layer, and the number attached to it will pull the entire policy conversation toward it.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
