Gates Says the World Needs About Two Decades to Absorb AI Disruption, and Governments Are Behind
Bill Gates told The Verge that the world needs to prepare for roughly two decades of AI disruption, and that the response so far has broadly not kept pace with how quickly the challenge is arriving. The remarks came in an interview with Decoder, published October 22, 2026. The Verge's interview is the most recent of a series of public statements Gates has made this autumn.
The series began in late August. In an essay, Gates wrote that even under the best circumstances, the transition to the new AI era will be one of the most turbulent periods in human history. He argued that a plan is needed to ensure that AI's benefits outweigh its harms. Gates Notes published the essay on August 26, the same day CNBC reported that he had warned governments are not adequately preparing for how AI could displace workers, strain social systems and create global risks.
In the Verge interview, Gates went further on the social side. He said AI disruption could be painful enough to split society in two, between people who are comfortable living near the technological frontier (the leading edge of what these systems can do) and people who reject it. That framing moves the question beyond employment statistics. It concerns whether a population can absorb a change of this speed while keeping its institutions, meaning the courts, schools, unions and public services that hold society together, intact.
His language on risk is blunt. Gates has called AI "the most dangerous thing that humans have ever gone near," and has said it "makes nuclear weapons look like nothing." He has also held a more conditional position. In an earlier conversation with The Verge, he said: "I believe AI is a beneficial thing, we just need to shape it in the right way." As he presents them, the two positions are not in conflict. The benefit is conditional on the shaping, and the shaping is the part he says is lagging.
On government readiness, Reuters reported on September 15 that Gates said governments worldwide are way behind on AI. He listed risks to jobs, addiction and attacks. A New York Times opinion piece titled "Bill Gates's Blunt Warning on A.I." on September 29 carried a more technical example. Gates said AI systems are finding bugs in code that humans had reviewed for more than 20 years without spotting a problem. For practitioners, that claim is the most open to direct testing, since it points to a concrete capability rather than a forecast.
The Gates Foundation is also part of the picture. Gates Notes says the foundation plans to spend its remaining $200 billion over 20 years. That is an intent and a schedule, not a completed programme.
In my view, the most useful question for engineers and policy staff is which of these claims can be checked over the next few years. Job displacement, the strain on social systems and the bug-finding claim each produce observable data. The two-decade horizon and the social split do not. Readers will weigh those differently, and the distinction matters when judging how much of the warning is forecast and how much is measurement.
The broader context here is that Gates is not arguing against the technology. Across the August essay, the Reuters-reported remarks and the Verge interview, his consistent argument is that adoption will outrun governance unless institutions move. If that gap closes, the same tools that make him uneasy could also shorten the time it takes to find and fix flaws in software that society already depends on. Whether the gap closes is the open question, and the coming decade will answer it more directly than any essay can.
