Godfather of AI Warns Mass Unemployment Is Coming. Musk and Gates Have Said the Same Thing.

Geoffrey Hinton isn't hedging anymore.

The British computer scientist who helped build the neural networks underlying modern AI, and who won the Nobel Prize in 2024 for that work, told Sen. Bernie Sanders (I-Vt.) in a November 2025 discussion at Georgetown University that mass unemployment from AI isn't a fringe worry. It's the likely outcome.

"It seems very likely to a large number of people that we will get massive unemployment caused by AI," Hinton said, according to Fortune.

His reasoning is blunt. Tech companies are pouring roughly a trillion dollars into data centers and chips. That money has to come from somewhere. Hinton's answer: selling businesses AI systems that do the work of human employees for a fraction of the cost.

"These guys are really betting on AI replacing a lot of workers," he said.

Hinton isn't out on his own limb here. He's essentially agreeing with the people running the companies building this technology.

Nvidia CEO Jensen Huang has predicted every job will be transformed by AI, possibly pushing society toward a four-day workweek. Bill Gates has said humans may soon not be needed "for most things." Elon Musk has gone further, predicting that in under 20 years most humans won't need to work at all.

What sets Hinton apart is his background. He spent decades inside the machine-learning research world before quitting Google in 2023 specifically to warn about AI's risks. His validation of the billionaires' predictions rather than dismissal of them as hype carries weight. He's told Fortune previously that he thinks the AI industry is driven less by scientific ambition than by short-term profit incentives, a sharper indictment than most tech executives level at their own industry.

There's a real tension worth naming. OpenAI, the company behind ChatGPT and arguably the face of this entire boom, isn't expected to turn a profit until at least 2030, according to HSBC estimates published in November 2025. HSBC estimated OpenAI may need more than $207 billion to sustain its growth.

That's the awkward part of the "AI will eliminate your job" narrative. The companies making that prediction are also the companies that need to justify enormous, currently unprofitable capital spending to investors. If AI replacing workers becomes the story that keeps valuations up, that's also the story that keeps the money flowing into Nvidia chips and hyperscale data centers.

None of that means Hinton is wrong. But it's a reason to treat predictions from people with a direct financial stake in AI adoption, including Musk, Gates, and Huang, as interested testimony, not neutral forecasting. Hinton, notably, has no comparable financial stake since leaving Google, which gives his warning more weight, though it doesn't make him infallible either.

Sen. Sanders has tried to put a figure on the disruption. In a report his office released in October 2025, partly built using estimates generated by ChatGPT itself, Sanders warned that close to 100 million U.S. jobs could be affected by AI automation.

That's an eyebrow-raising number, and it's worth being honest about its shakiness. Using a chatbot to help estimate the job losses that chatbots might cause is not exactly a rigorous labor-economics methodology, and Sanders' office has not published a detailed jobs-by-sector breakdown alongside it that would let outside economists check the math.

Hinton, to his credit, doesn't pretend to have a crystal ball. He compared predicting AI's trajectory to driving in fog.

"You can see clearly for 100 yards and at 200 yards you can see nothing," he told Sanders. "We can see clearly for a year or two, but 10 years out, we have no idea what's going to happen."

He also acknowledged AI will create new jobs, which is what happened with past waves of automation, from tractors to computers to the internet. His disagreement with more optimistic tech leaders isn't over whether new jobs will appear. It's over whether there will be enough of them to offset what's eliminated. He doesn't think there will be.

Nobody, including Hinton, has a solid number for net job loss versus job creation over the next decade. There's no consensus timeline, no agreed-upon policy response, and no functioning federal framework for retraining workers at the scale Sanders is describing.

What is measurable right now is the spending. Trillions of dollars are flowing into AI infrastructure on the bet that labor substitution, not just labor augmentation, is where the return comes from. Whether that bet pays off, and who absorbs the cost if American workers are displaced faster than new roles appear, remains the open question neither the tech industry nor Congress has answered.