Congressman Lieu Proposes Mandatory AI Kill Switch Legislation

Bipartisan push for AI safety controls

Congressman Ted Lieu is advancing legislation that would mandate "kill switch" capabilities for artificial intelligence systems, allowing humans to shut down AI programs that behave unexpectedly or dangerously. The California Democrat, co-sponsoring the bipartisan AI Kill Switch Act, outlined the proposal's core principle in an interview with CNN's Erin Burnett.

"It should be the first principle of AI safety that humans must be able to turn off AI systems and AI agents," Lieu stated on Erin Burnett Out Front.

The legislative push comes as concerns about AI safety and control continue to intensify across government, industry, and the public. The proposed requirement would establish human override as a fundamental safeguard in AI system design, ensuring operators maintain the ability to deactivate programs regardless of their operational state.

Why it matters

As AI systems grow more autonomous and are deployed in critical infrastructure, healthcare, and national security applications, the ability to immediately shut down malfunctioning or compromised systems becomes a fundamental safety requirement. Mandating kill switches addresses a key vulnerability: AI systems that could resist shutdown attempts or operate beyond human control. This legislative approach signals growing bipartisan recognition that voluntary industry standards may be insufficient for managing AI risks.

Technical and policy implications

The kill switch requirement would likely affect AI developers across sectors, from large language models to autonomous systems. Implementation details remain unclear, including how the legislation would define "AI systems and AI agents" subject to the mandate, what technical standards would govern kill switch design, and how enforcement would work across different deployment contexts.

The bipartisan nature of the bill suggests potential viability in a divided Congress, though the legislation's specific provisions and timeline were not detailed in Lieu's remarks. The proposal joins a growing body of AI safety legislation at federal and state levels as policymakers grapple with governance frameworks for rapidly advancing technology.

The emphasis on human control reflects broader debates in AI safety research about maintaining meaningful human oversight as systems become more capable and autonomous. Critics of current AI development practices have warned that systems could become difficult to control or align with human values without built-in safety mechanisms.

Details of the AI Kill Switch Act were first reported by CNN during the Erin Burnett Out Front segment.