A new bill in Congress would give the Department of Homeland Security the authority to order the throttling or complete shutdown of frontier artificial intelligence systems. Companies that fail to comply could face fines of up to $20 million per day.
What the AI Kill Switch Act Would Do
The legislation, known as the AI Kill Switch Act, would empower DHS to intervene when it determines a frontier AI system poses a risk to national security or public safety. The department could issue an order requiring the system's operator to reduce its capabilities or shut it down entirely. Noncompliance triggers a daily fine that escalates quickly, potentially reaching hundreds of millions of dollars over a few weeks.
The Challenge of Defining Frontier AI
The bill does not define what qualifies as a frontier AI system. That leaves DHS and the courts to interpret which models fall under the law. Critics argue the vagueness could lead to overreach, while supporters say it gives regulators the flexibility needed to keep pace with rapid technological change. The term is generally understood to refer to the most advanced and capable AI models, but no official definition exists in the text.
Potential Impact on AI Development
If enacted, the act would mark a significant expansion of federal oversight over AI. Companies developing cutting-edge models would need to prepare for possible government intervention. The threat of daily fines could push firms to build in compliance mechanisms from the start, or to slow down deployment of new systems. The bill does not specify a process for appeal, raising questions about due process for affected companies.
The legislation is in its early stages. Lawmakers have not yet scheduled hearings or a vote. The proposal is likely to face debate over the balance between innovation and security.
