Bill Gates’ warning did not arrive as a blanket rejection of artificial intelligence. It marked a sharper turn from one of the technology sector’s best-known optimists: AI, he now argues, may be transformative, but its rollout is racing ahead of the world’s ability to contain its worst uses.

In an August essay, Gates said there was no credible plan for deploying AI safely amid the social, political and economic disruption ahead. He wrote that he would support a global slowdown if one could be carried out credibly.¹

That concern intensified in a clip released Thursday ahead of his full Meet the Press interview. Gates warned that “AI is certainly powerful enough to drive events that cause a billion deaths,” focusing on the prospect that bad actors could gain access to sophisticated tools.¹ The concern is not purely hypothetical in the industry’s own safety debate: Anthropic has described actors using AI to plan drone attacks and conduct dangerous biological research.¹

Gates’ prescription is direct: government cannot remain a spectator. “You need law enforcement and the politicians to get into the discussion about what safeguards and monitoring look like,” he said.¹ His argument aligns with calls from Anthropic chief executive Dario Amodei for regulation to pace frontier AI development, though Washington has shown little appetite for a broad slowdown.¹

The opposing instinct is visible in a retweet by David Sacks, who amplified a comparison between banning AI and a Chinese emperor’s alleged ban on shipbuilding—a warning that technological restraint can leave societies behind.David Sacks (DavidSacks)² Microsoft chief Satya Nadella has occupied a middle ground, urging companies to stop when they see a “show-stopper,” rather than endorsing sweeping external controls.¹

The divide is therefore not over whether AI needs guardrails. It is over who applies them, how fast, and whether fear of catastrophe justifies slowing a strategic technology before catastrophe arrives.