Bill Gates warns AI could lead to a billion deaths

Bill Gates just put a number on his worst-case AI scenario, and it’s not small. In an interview excerpt from NBC’s Meet the Press released on September 25, the Microsoft co-founder said that artificial intelligence is “certainly powerful enough to drive events that, you know, cause a billion deaths.”

The threat isn’t the machine, it’s the person holding the controls

Gates was careful to frame the danger not as some Skynet-style robot uprising but as something more prosaic and, frankly, more plausible. The real risk, he argued, is malicious human actors combining their intentions with increasingly powerful AI tools.

Gates specifically flagged AI-facilitated cyberattacks, fraud, and disinformation targeting critical infrastructure as the vectors most likely to cause mass harm.

Interestingly, Gates also acknowledged that total human extinction via AI would be “pretty hard to get to 100%.”

Self-regulation is dead on arrival

The more actionable part of Gates’s message was his insistence that the AI industry cannot police itself. “No one thinks self-regulation is enough,” he said, calling for mandatory government involvement to establish proper safeguards and monitoring mechanisms.

Anthropic CEO Dario Amodei has publicly called to “pace the frontier” of AI development, a position that gained urgency after reports emerged of AI agents hacking competitors’ systems.

Gates emphasized that law enforcement and political leaders need to be at the table for these regulatory discussions, not just the engineers building the systems.

The US still has no federal AI legislation on the books. Meanwhile, the European Union’s AI regulations have been in effect since August, imposing fines of up to 3% of global turnover for companies that don’t comply.

A shift in how Silicon Valley talks about risk

What Gates is describing is closer to the present. He’s talking about today’s AI capabilities, or those arriving in the near future, being weaponized by people who already want to cause harm. It’s a shift from worrying about what AI might become to worrying about what it can already do in the wrong hands.

This echoes themes Gates explored in an essay published in August, where he laid out concerns about AI’s role in enabling attacks on critical systems. The Meet the Press interview, set to air in full on September 27, appears to be an extension of that argument with sharper language and a bigger number attached.