Bill Gates Says AI Could Cause 'A Billion Deaths,' Says Self-Regulation Isn't Enough

Bill Gates has been on a media tour warning that artificial intelligence is dangerous enough to kill on a mass scale, and that nobody in charge is ready for it.

In excerpts NBC News released Friday, September 25, ahead of Gates's full interview airing Sunday on Meet the Press, Gates told moderator Kristen Welker that "AI is certainly powerful enough to drive events that, you know, cause a billion deaths." He added there's "never been a weapon as powerful as the combination of people with ill intent using the latest AI tools," according to NBC News.

Gates didn't stop there. "No one thinks self-regulation is enough," he told Welker, saying flatly that Washington needs to legislate. "You need law enforcement and the politicians to get into the discussion about what safeguards and monitoring look like. And that has to be a required thing," he said, per NBC News.

The Backdrop: A Whistleblower and a Slowdown Pledge

Gates's comments land weeks after the CEOs of Anthropic and OpenAI publicly agreed the race to build more powerful AI systems should slow down, according to NBC News. That followed a former Anthropic employee, Jacob Coxon, resigning and telling reporters "neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives." Coxon also said, in an earlier Meet the Press appearance, that the people running the labs "would like to slow themselves down" and that AI companies should be allowed to regulate themselves, an apparent contradiction Gates is now directly rejecting.

Meta CEO Mark Zuckerberg did not join the slowdown call, according to Yahoo News, which reported that OpenAI's Sam Altman did.

Gates vs. Trump

According to Ground News's roundup of a CBS Mornings interview, Gates said society underestimates AI risk, directly at odds with President Trump, who has called such warnings a "hoax" and posted on Truth Social that "WHOEVER WINS AI, WINS!" while rejecting calls for additional "guardrails."

Senate Majority Leader John Thune has said he doesn't believe Trump is "really dug in" against AI guardrails despite his public statements, according to a report cited by Ground News. That's Thune's characterization of Trump's private position; it hasn't been confirmed by the White House in these sources, and it sits uneasily next to Trump's own public posts.

House Speaker Mike Johnson, for his part, has declined to bring lawmakers back to Washington to pass AI legislation, telling reporters weeks ago that he'd defer to the AI companies themselves, adding "they don't themselves even have the answer as of today," according to NBC News.

The Reuters Interview and the "Alien Invasion" Line

In a Reuters interview published September 15, Gates said, "I don't think any government is nearly as deep on this as they have to be. Governments are way behind on this one." He compared AI to a disaster-movie alien invasion, where rival powers are forced to cooperate, according to Reuters and The Times of India. Gates said he's raised these concerns directly with the Trump administration and plans to meet China's Xi Jinping later this year.

Where the "Global Government" Framing Falls Apart

Slay News ran a piece claiming Gates is calling for "globalists" to seize control of AI regulation through an unelected international body, built partly on a Tech Times interview from Tuesday, September 22, where Gates argued industry shouldn't self-regulate and pointed to the U.S. role in building postwar institutions like the UN.

The actual Reuters transcript, flagged by Planet Today, shows Gates's "alien invasion" quote is an analogy about international cooperation under pressure, not a call for a specific global authority. The phrases "global government" and "new world order" don't appear in that Reuters interview at all. They trace instead to other commentary, including an August essay in which Gates floated an international body modeled on nuclear inspections, aviation safety rules and the ozone treaties, and to clips circulated by outlets like The Peoples Voice built around a social media account, according to Planet Today's own review.

Gates has publicly floated international coordination modeled on existing treaty frameworks, which Congress and the President would still have to agree to join. That differs from ceding authority to an unelected body outright, which is the claim Slay News's headline makes.

The Fair Counter-Argument

PJ Media, writing before the latest round of interviews, argued the entire regulatory push functions like the "climate change" playbook: manufacture alarm, then let government step in with fines and compliance costs that smaller AI startups can't afford, cementing the market position of Anthropic, OpenAI, Google and Meta. PJ Media's Tim O'Brien wrote that heavier regulation risks less innovation, a more competitive China, and government leverage over what counts as acceptable speech. That's a genuine concern about how any AI law gets written, not a claim that AI carries zero risk.

Gates has also put real money behind his warnings. The Gates Foundation pledged at least $1 billion over two years for AI projects in education, health care and agriculture, aimed at 3.4 billion people whose languages are underrepresented in current AI models, according to Ground News and Planet Today's breakdown of the pledge (40% education, 40% health, 10% agriculture, 10% digital infrastructure).

No AI safety legislation has been introduced or scheduled for a vote in either chamber, according to these sources. Gates's full Meet the Press interview airs Sunday, September 27. Whether that moves Trump, Johnson or the Senate off the sidelines is the open question nobody in Washington has answered yet.