Bill Gates: a kill switch alone cannot prevent AI harms

‘We do need to monitor any sophisticated model and record exactly what's being done’

EXCERPT

GATES: "The kill switches, that, that's kind of a, a weird thing. Uh, it's not enough to have a kill switch. They, they, they... We're not yet at the point where they, they, they autonomously grab computers and, you know, can't be shut down. So th- that discussion is just sort of shows how non-technical various people are. Uh, you know, I'd, I n- would never wanna say I'm against a kill switch, but you, you don't... When you, when you're trying to moderate the, the bad behavior, you need insight and records of what's being done. So the kill switches alone would not be, not prevent these tragedies. What would prevent tragedies from happening due to AI specifically? Well, there... I, I wish it was easier to distinguish good intent, like I wanna design a new drug for a disease versus, uh, negative intent, which is I wanna design a tool of bioterrorism. Mm. Because those are so close, we do need to monitor any sophisticated model and record exactly what's being done. The United States ought to make sure this is done. Uh, China will wanna do this. Uh, usually people expect the US to be very concrete about what this looks like. Um, and you'll hear from the industry it's, it's a very doable thing. It would not slow them down dramatically. And so it's this, these safeguards and monitoring are separate from would we be better off slowing down."

Source: NBC News · Sep 27, 2026 · 109s clip