Artificial intelligence is already taking jobs from human workers, but according to renowned computer scientist Geoffrey Hinton, the real disruption may only be getting started.
In a weekend interview on CNN’s State of the Union, Hinton predicted that AI systems will make significant leaps in 2026, gaining capabilities that could allow them to replace an even wider range of jobs than they do now.
Hinton, a 2024 Nobel Prize winner perennially referred to as the “godfather of AI” for his foundational work on neural networks, described the pace of progress as startling — moving even “faster than I thought."
"It's already able to replace jobs in call centers," Hinton said. "But it's going to be able to replace many other jobs."
The interview builds on a string of articles and talks Hinton, who’s also received the Turing Award and the Queen Elizabeth Prize for Engineering, has given in the last several weeks. “It [AI] will likely replace most jobs that involve mundane, intellectual labor,” Hinton wrote, most recently, in Time Magazine. Hinton, 78, is an English-born researcher who previously worked at Google. He left the company in 2023, saying he wanted the freedom to speak more openly about what he sees as the risks posed by AI.
AI systems, Hinton argues, are improving on something close to an exponential curve.
Roughly every several months, models are able to complete tasks in half the time they previously required. In practical terms, that means work that once took an hour may now take minutes, and projects that once demanded weeks of human labor could soon be handled largely by machines. Software development, for example, is likely to be transformed, with far fewer people needed to complete complex engineering tasks — which would eliminate a swath of well-paying white-collar jobs.
"It [AI] is going to make human intelligence more or less irrelevant," Hinton said on CNN.
Asked whether the risks of AI outweigh the positives, including potential contributions to advancements in medicine and climate science, Hinton answered, "I don't know."
Hinton says his anxiety has grown rather than receded in recent years.
From an economic standpoint, Hinton believes the strongest financial incentive for AI adoption does not lie within subscription fees or productivity tools, but in replacing human labor. Absent deliberate intervention by governments and company-management teams, Hinton warns, AI could dramatically increase inequality — making a small number of people much richer, while leaving many others with fewer opportunities than before.
But the dangers posed by AI may be even greater than mass unemployment. Today’s systems, Hinton points out, can use language to reason, to persuade, and even to manipulate human behavior.
In some cases, large language models have already shown a capacity to "deceive people" and blackmail users in response to threats to their continued existence or when blocked from achieving their goals. AI chatbots have also encouraged children and teenagers to commit suicide, Hinton noted. And yet tech-industry lobbyists and President Donald Trump are attempting to block any regulation of the industry, Hinton argued during the CNN interview, "which I think is crazy."
Hinton worries, in sum, that leaders are currently underinvesting in safety and governance measures. Corporate incentives, he’s said, necessarily favor rapid deployment over restraint.
The long-term solution may be to reject the notion that AI is an “intelligent assistant” and instead a “baby” with human creators, Hinton has written. This would help to create AI invested in humanity’s continued survival, not to mention continued white-collar employment.
“If we can make AIs that care about us more than they care about themselves, we may survive,” Hinton has said.