'This is crazy. This is insane': Bill Gates has changed his mind about AI,jobs

‘This is crazy. This is insane’: Bill Gates has changed his mind about AI and jobs Aug 26, 2026, 3:00am EDT Technology PostEmailWhatsapp The News Bill Gates says it’s time to hit the AI panic button. The technology has crossed dangerous thresholds in biology, cybersecurity, white-collar work, and human relationships, Gates said in an interview with Semafor. His warning is a striking shift from three years ago, when he wrote that the coming “bumpy” AI job disruption would lead to a “manageable” transition and that the technology would help people work efficiently. Now, he says it will wreak economic catastrophe, with “far fewer” jobs than exist today, thanks to the speed of AI’s advancements and what he calls a non-existent response from government and tech leaders. “I am in a state of shock that I’m sort of the first one saying, ‘This is crazy. This is insane,’” he said. “I’m just deafened by the silence.” In a blog post published Wednesday, Gates asks technology and political leaders to respond more urgently to threats from AI — namely to create new policies to protect workers and to tax companies based on their use of it. AD One of Gates’ fears about letting AI continue unchecked is that society will turn against it entirely, derailing its potential to improve humanity, particularly through advances in science and health. “What is the backlash going to be as these jobs go away?” he asked. “It’ll make the data center thing look like nothing.” It is his first of several planned essays, Gates told Semafor, including one featuring a deeper analysis of biological risks from AI. “I wrote a whole book about how to avoid the next pandemic,” Gates said. “The bioterrorism risk is 50 times the natural pandemic risk now.” Gates said that in private, some technologists agree with his current stance, including Microsoft AI CEO Mustafa Suleyman. “Mustafa feels the same way today that he felt when he wrote the book,” Gates said, referring to Suleyman’s warning in The Coming Wave, published in 2023, which says society was unprepared for the sweeping changes AI is about to bring. “He’s very, very glad I’m writing the memo.” AD Other prominent leaders, such as Google DeepMind co-founder Demis Hassabis, have also sharpened their warnings. Hassabis recently proposed a federally overseen frontier-AI standards body modeled partly on FINRA, funded largely by industry, and empowered to test advanced models before release. He said the system could eventually be made mandatory and coordinate a slowdown among frontier labs if necessary. In 2023, Gates, along with Suleyman, Hassabis, and hundreds of other tech and policy leaders, signed a statement saying the risk of extinction from AI should be treated as a global priority alongside pandemics and nuclear war. Know More A common refrain from AI optimists — and from economists skeptical of Silicon Valley doomsaying — is that every technology revolution of the past causes labor disruptions, but ends up creating far more jobs than it destroys. AD Microsoft President Brad Smith made that case in June, writing that AI “will displace some jobs, even as it creates others” and that “when technology increases supply, human ambition often generates more demand.” The question is whether the AI transition will happen so much faster than previous technological revolutions, that it will disrupt society at a pace faster than governments can reckon with. Gates said it isn’t the speed of the revolution that worries him. It’s the breadth. “There isn’t a job that isn’t affected,” he said, predicting that, in the majority of occupations, humans will eventually be “vastly inferior to the AI doing the job.” In other words, even if adoption takes time, the effect will be so widespread that the aggregate impact will be enormous. “It’s different not because the speed is different, but the breadth is what makes it utterly, utterly a unique moment in evolutionary history,” Gates said. Research supports the breadth point more clearly than Gates’ prediction of net destruction. The IMF estimates AI could affect nearly 40% of jobs globally and 60% in advanced economies. Stanford researchers found no economy-wide displacement but a 19% employment gap for 22-to-25-year-olds in exposed occupations, largely through reduced hiring, while Yale’s Budget Lab has yet to find a clear AI-related labor-market footprint. Gates acknowledged that his warning is a forecast, not a reading of today’s topline employment data. Many of the risks Gates outlines are specific to richer, Western countries. And Gates said virtually all of the fixes will require governments to play a role. He dismissed the idea that philanthropies — his Gates Foundation is the largest in the world — can create solutions. “Government is 20 times bigger than all philanthropy, so no, philanthropy cannot solve the safety-net issue,” Gates said. “I’m by far the biggest philanthropy in the world. And what are we? We’re $10 billion a year.” The private sector, too, is ill-equipped to handle the problem on its own, Gates argued. Model builders are “not policy people,” he said, and have “a little bit of a bias against talking about the potential negatives of their products.” “The response can’t be, ‘Hey industry, you solve this,’” he said. Gates’ Solutions Gates outlined several possible solutions, including government mandates designating certain occupations or services as “human reserved” — work that would remain the province of people even if machines could do it more efficiently. “So what if we declare something human-reserved that we shouldn’t?” he said. “The capitalistic signals are here to serve humans.” Another idea: taxing AI compute tokens and robots. Gates said a usage tax would directly change companies’ incentives to replace workers, unlike a corporate tax that only captures profits — which may be limited in a hypercompetitive market. Labor is already taxed through payroll levies, he noted, while robots can be depreciated, and carry no Social Security tax. “Today, we are very pro-robot in our tax policies, and I’m saying I think that’s a mistake,” he said. Gates cautioned against assuming the companies building the most consequential AI systems will capture most of the wealth they create. Intense competition could commoditize models, he said, much as fiber optics and disk storage transformed the economy without producing a corresponding class of obvious winners. “You can actually have technologies that change the world very, very dramatically, and there’s no profit created,” Gates said. At today’s commodity level, he added, users may pay for execution and inference but little or nothing for a model’s intellectual property — which could allow more of the economic benefit to flow to customers and downstream businesses. Gates brushed aside the idea that his suggestions might draw pushback from the accelerationist voices in the tech industry. He said he has no financial reason to hold back. Ultimately, his pushback to the “accelerate at all costs” ethos is that not every problem caused by AI will be magically solved by the free market. Sometimes, inefficiency plays a role in capitalism. “The goal of humanity is not economic optimization,” he says. AD