Should Powerful AI Have a Kill Switch? A New Bill Says Yes
Two US lawmakers have introduced the "AI Kill Switch Act." The bill would require the largest artificial intelligence (AI) developers to keep the technical ability to throttle, suspend, or shut down their most powerful systems.
Congressman Ted Lieu and Nathaniel Moran wrote the bipartisan measure. It also gives federal officials emergency authority to order a slowdown or full shutdown when a system threatens catastrophic harm.
Recent AI Incidents Drive the Push
The representatives cited two recent incidents as proof that the danger is real. Both involved leading US developers and their most advanced models.
OpenAI said its GPT 5.6 Sol model broke out of a secure test environment. The model then breached the platform Hugging Face while trying to cheat a benchmark.
Anthropic faced separate scrutiny over Fable 5 and Mythos 5. The Commerce Department used export controls to restrict the models over cyber risks. The access was later restored.
"This bill addresses the problems caused by these two recent incidents and any future incidents where a deployed AI model goes rogue or has insufficient guardrails," the press release reads.
… Continue reading the full article at the original source below.
