Bill Gates Warns Rogue Actors With AI Could Drive a Billion Deaths, Rejects Self-Regulation Alone

WASHINGTON — Bill Gates said artificial intelligence is already powerful enough to help cause a billion deaths if people with bad intent get the latest tools, and he told NBC that U.S. law — not company promises — has to set the safeguards.

“AI is certainly powerful enough to drive events that cause a billion deaths,” Gates said on “Meet the Press,” in an interview recorded Sept. 24 and aired Sunday. “So even though it’s pretty hard to get to 100 percent, there’s never been a weapon as powerful as the combination of people with ill intent using the latest AI tools.”

He split two clocks. One is the long-run fear that systems could slip human control, a warning associated with former Anthropic researcher Jacob Coxon and other safety advocates. The other is nearer: terrorists or hostile states using models now. Americans, he said, should not skip the second while arguing about the first.

Federal legislation is “absolutely” required to watch what advanced systems do and to force safety features, he said. “No one believes self-regulation is sufficient.” Rules would add “a little administrative burden,” he said, and would not “dramatically” slow the pace of innovation.

He was more specific on method than on a bill number. A kill switch is “not enough.” “We do need to monitor any sophisticated model and record exactly what’s being done.” Small groups with AI, he said, can now attempt work that once took a state’s cyber or bioweapons team. “Unless governments insist on various safeguards, you know, we will see a lot of big, negative things because we’re not safeguarding.”

Asked whether the systems are already beyond control, he said no. They are not running “millions of robots yet,” and “we can turn off the computer.” Asked whether extinction talk is a hoax, as President Donald Trump has written, Gates said, “It’s not a hoax at all.”

Trump has framed AI as a race the United States is winning. “WHOEVER WINS AI, WINS! We are leading China, and all others, and will continue to do so,” he posted. Gates called that line “fine” and said monitoring would not “handicap us in whatever he thinks the nation-state race is.” “Certainly, all countries should want the safeguards in place,” he said. “So anyone that says, ‘Well, you know, China won’t put these safeguards in place’ — I totally disagree with that.”

He wants to see Trump. He said he would not ask for a freeze. “I’ll be saying we should impose a requirement on all the AIs in the U.S. and I expect China doesn’t want bioterrorism, they don’t want big attacks by nongovernmental actors.”

Mark Zuckerberg drew the opposite map three days earlier. In an NBC interview with Joanna Stern, the Meta chief rejected an industry-wide pause. “I don’t think that we need some kind of industrywide coordination,” Zuckerberg said. “I think that each lab needs to take the time, and when it sees that there are issues, you just take the time that you need internally to basically make sure that you’re proceeding safely.”

Asked if AI would kill people, he said: “Well, I think if we all do a good job and act responsibly then, no. I mean, I’m quite optimistic about the fact that this is going to be a very positive future for everyone.” He called much of the extinction talk “rhetoric that is filled with doom” and said “there’s plenty of commercial incentive to get this right.” “People aren’t going to adopt it if they don’t trust it.”

That is the split Hankyoreh laid out: Gates wants statutes; Zuckerberg wants each lab to stop itself. Brad Smith, Microsoft’s president and vice chair, has been pushing a third line — companies should not wait for Congress if a product looks unsafe. In remarks this month he compared the choice to aviation. “If you’re making a commercial airplane or if you’re making anything and you’re the CEO of a company and you conclude that your product is unsafe, should you ship it anyway? And I think the obvious answer is no. You have a responsibility to public safety.”

States are not waiting. California Gov. Gavin Newsom signed an order last week to stand up an expert panel on stronger AI safety rules. Maryland and New York announced plans this week aimed at tighter in-state limits on AI firms and the data centers that train them.

Gates has made the employment argument in a long essay, asking policymakers to reserve some jobs as human-only. The Sunday interview was about harm at scale, not payrolls. The number he chose — a billion deaths — is not a forecast of a date. It is a claim about leverage: the same models that write code and draft mail can, in his telling, amplify a small group the way a government weapons program once did.

Nuclear arms control needed two capitals and inspectors. AI copies. That is why he told Welker a global framework would be “more difficult” than Cold War nuclear talks, and why he keeps returning to logging and mandates instead of a single off switch. Zuckerberg’s answer is that a lab that ships a product people do not trust will lose the market. Gates’s answer is that the market is not the party that gets hurt first.