Rep. Ted Lieu warns that rapid advances in artificial intelligence are outpacing existing safeguards, arguing that Washington must establish an emergency “kill switch” for powerful systems that pose imminent risks. Citing the potential for unintended behavior, misuse, and cascading failures across critical infrastructure, he urges Congress to empower regulators with authority akin to aviation or market circuit breakers—backed by standards, audits, and incident reporting. The proposal favors targeted oversight of high-risk deployments rather than a blanket slowdown, with requirements for red-teaming, monitoring, and shutdown protocols. Lieu frames the approach as pro-innovation but safety-first, calling for bipartisan action to ensure AI benefits don’t come at the expense of public security.
Related articles:
AI Risk Management Framework (NIST)
Open Letter: Pause Giant AI Experiments
