U.S. Lawmakers Introduce Bipartisan Bill Requiring Kill Switch for the Most Powerful AI Systems

My take: A June 2026 survey found that 86% of likely U.S. voters support requiring the most powerful AI models to have an emergency shutdown mechanism. The number barely moves across party lines: 88% of Democrats, 86% of independents, and 83% of Republicans. That level of bipartisan consensus is rare on any public policy issue.

That agreement is what prompted Representatives Ted Lieu (D-CA) and Nathaniel Moran (R-TX) to introduce the AI Kill Switch Act. The bill would require companies whose AI models consumed more than $100 million in compute to maintain the technical capability to throttle or shut down those systems in an emergency. Non-compliance would cost up to $2 million per day, rising to $20 million for violating an emergency order.

For anyone building with AI or making adoption decisions, this kind of regulatory framework is not bad news: it is a signal that governments are trying to sustain public trust in the technology over the long term. Without that trust, access to these tools gets harder for everyone.

If your company uses third-party AI models, do you have a clear plan for responding if one of those models causes an unexpected incident?