Gates says AI safeguards would not hamper US against China

Bill Gates said Sunday that government safeguards against catastrophic AI risks would not hamper the United States in its competition with China. His comments, on NBC News's "Meet the Press," contrast with President Donald Trump's more hands-off approach, Bloomberg and NBC News reported.

Reaching a global AI agreement will be "more difficult" than Cold War-era nuclear arms negotiations, Gates told NBC. Safeguards and monitoring would not "handicap us in whatever [Trump] thinks the nation-state race is," he said. He added that the US should lead a global framework, as it has on aviation safety and nuclear weapons.

Gates said China would likely restrict its own dangerous AI models if the US moved first, according to NBC News. He named preventing people with "bad intent" from using AI for cyberattacks, bioterrorism or fraud as his most immediate concern.

Gates called for mandatory internal monitoring at AI labs, saying a "kill switch" alone would not prevent a catastrophic failure. He separately said his foundation is working to bring AI language support for African and other non-English languages up to English-level quality. He set a five-year target, NBC News reported.