Bill Gates Says Global AI Framework May Be Harder Than Cold War Nuclear Talks and Is Not a Hoax

WASHINGTON — Bill Gates said Sunday that getting countries to agree on a global framework for artificial intelligence could be harder than Cold War talks that limited nuclear weapons, and he rejected President Donald Trump's claim that catastrophic AI warnings are a hoax.

"This is more difficult than that was," Gates told Kristen Welker on NBC's "Meet the Press" after she asked whether U.S.-Soviet nuclear limits could be repeated with China on AI.

He said the United States should lead that work the way it once led other postwar bargains. "Since World War II, the U.S. should be very proud that when there have been global challenges, we, in a way that was benign, where we had friends and allies coming along with us, you know, aviation safety, ozone layer, nuclear weapons, a lot of things where we played the fair actor, not aggrandizing whatever our unique strengths were," Gates said. "And AI is one we're going to have to go back to this, 'Hey, we're all in — humanity is all in this together.'"

Trump has opposed new federal guardrails on model builders. On Truth Social he wrote, "WHOEVER WINS AI, WINS! We are leading China, and all others, and will continue to do so." He has also called claims that AI could destroy humanity a "HOAX."

Gates said Trump's competitive line was "fine," then predicted a shift. "What he said is fine, but I think after he hears from me and other people, including that this does not handicap us in whatever he thinks the nation-state race is, that, you know, we will come to a consensus."

On the hoax charge he was direct. "It's not a hoax at all," he said. "I hope that, given my life's work in the field, my saying how unique and different this is and how concerned I am will add to what he's hearing from other people."

He wants a meeting with Trump. He told Welker he would not argue for a development 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."

Asked whether AI could already be out of control, Gates said no. It is not running "millions of robots yet," he said, and "we can turn off the computer." He still said it "absolutely" needs government rules plus industry rules. A kill switch alone is "not enough." "We do need to monitor any sophisticated model and record exactly what's being done."

On harm, he stopped short of saying AI would end the species and then set a high number. "AI is certainly powerful enough to drive events that cause a billion deaths," he said. "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 tied that to a change in who can do damage. "Historically, countries, you know, had cyberattack teams and they could think of things like bioterror weapons. Now, small groups with AI can do what only the biggest countries could do. So it's really the bad intent plus AI that looms as a major threat. Unless governments insist on various safeguards, you know, we will see a lot of big, negative things because we're not safeguarding."

He said industry talk about risk was "great," but the next problem is people "either discounting this or just thinking of this as a competition." "The problem is that the rest of the world is either discounting this or just thinking of this as a competition. And it's very scary that we don't have all of the thinkers saying, 'How do we want to shape this?'"

Because the debate is "cast in this nationalistic, you know, us versus them type thing, you have people who are saying, 'Well, let's just go full-speed ahead.' It's completely irresponsible not to require every AI to have these safeguards and monitoring capabilities. And it only is a question of do you do it now?"

"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 also said self-regulation is not enough: "No one believes self-regulation is sufficient." Law enforcement and elected officials, he said, have to help design mandatory oversight.

The interview landed after similar public pressure from Anthropic chief Dario Amodei, OpenAI chief Sam Altman and Tesla chief Elon Musk. Trump's answer has been that existing guardrails are enough and that slowing U.S. labs helps Beijing.

Gates's nuclear comparison is the line that travels. Nuclear talks were between states that already possessed the weapons and could count warheads. AI is software that copies, ships and improves in private labs and on open models. That is why he called it harder. It is also why a bilateral U.S.-China handshake would not, by itself, cover the small groups he named.

He is not arguing that the computers are already loose. He is arguing that waiting until they are is the mistake. The switch still works, he said. The record of what a model does does not exist unless someone requires it. That is the gap he wants Washington to close without treating the requirement as a forfeit to China.