An Austrian-born economist who has spent decades studying the intersection of artificial intelligence and economic transformation is sounding a stark warning: Within a few years, AI will be capable of performing virtually every intellectual task humans currently do for a living — and the consequences for the labor market could be unlike anything in recorded history.
Anton Korinek, born in 1978 in Reichenau an der Rax and now a professor at the University of Virginia, has been working on the topic of artificial intelligence since the 1990s. In a wide-ranging interview with German newspaper Welt, she wades into a dark realm, including “a risk of mass misery” for humanity.
Asked how many people will lose their jobs to AI in the coming years, Korinek was careful but unsparing.
“There is a tremendous amount of uncertainty about this. There are no models for this because something like this has never happened before,” he said. “But what we can be sure about is that in a few years AI will be able to perform virtually all intellectual activities.”
Korinek is of course far from the only one sounding the alarm, with even the heads of the top AI labs in the world stating that the implications for the global labor market are dire. For example, Anthropic CEO Dario Amodei recently warned that AI could “wipe out” jobs across industries, while stating that “unusually painful” job cuts are ahead.
As has been seen almost ad nauseam since the rise of AI, the debate is whether AI will create new jobs or lead to a permanent replacement of human labor and input. Korinek appears to side with the latter view. He pushed back against the commonly held reassurance that new technologies always create new jobs to replace the ones they destroy.
“First of all, new technologies have not ‘always’ created new jobs, but in the past 250 years, i.e., since the industrial revolution, it is not the technologies that have created new jobs,” he said. “Work was always valuable and irreplaceable. At the same time, our economy continued to grow, and the population continued to grow. That’s why there were more and more jobs — not because of any technology.”
And when his interviewer pointed out that technologies have not yet led to mass unemployment, Korinek offered a pointed historical corrective: “The spinners and weavers who became unemployed as a result of the industrial revolution would disagree with you.”
Anthropic CEO Amodei would likely agree with this take, as he recently stated that the technology is not replacing a single job but “acting as a general labor substitute for humans.”
The fundamental question AI poses, he argues, is one that previous waves of automation never truly raised.
“Will the value of work as a whole continue to rise or will it fall if AI and later robots can perform all work?” he asked. “Until now, the answer has always been that automation only destroys a small portion of jobs. In the next few years, however, there is a risk that all the jobs that can be done in front of a laptop will be lost.”
He added that from there, robotics will disrupt nearly every other job.
When asked by the Welt journalist, Korinek said that AI could likely do his job, including the entire interview, to a reasonably satisfactory degree. Even if shortcomings remain, he noted that the models are quickly advancing every three months.
When the journalist asked how he would “buy vegetables for my family?” stating that”food won’t suddenly be free,” Korinek said that the cost of vegetables would even increase, since AI has such high energy demands.
Korinek said that the question about how people get by in such a scenario has dominated his thoughts.
“That’s exactly what I’ve been thinking about for 10 years,” he said. “There is a risk of mass misery.”
Korinek says he maintains a list of 10 to 12 professional areas that will remain after AI has transformed the economy — though he concedes that half of those are only temporary refuges. What survives in the long run, he argues, comes down to one thing: human presence and personality.
“In the long term, all that remains are all performative professions. Athletes, actors, and so on. Because personality matters — no one is impressed when a robot runs faster than an athlete.” Religious personnel — priests, rabbis — also make his list, as do people whose job it is to monitor AI systems. “If — but that’s a big ‘If’! — people keep control,” he added.
On the question of artistic creativity, Korinek offered one of the interview’s most provocative exchanges. Asked whether AI could write something like King Lear, he reframed the question entirely: “Can a neural network of neurons write ‘King Lear’? The answer is obviously yes, because a neural network wrote ‘King Lear’.”
Essentially, Korinek was implying that Shakespeare was himself a biological neural network. The implication, that the distinction between human and machine creativity may be less absolute than we assume, was not pursued further by the interviewer.
Finally, the conversation turned darker when Korinek addressed the risk of AI systems that surpass human intelligence entirely. Drawing on a vivid analogy, he noted that “no one has the slightest clue what happens when we create something significantly smarter than us.”
He compared the challenge of humans controlling superintelligent AI to toddlers in a kindergarten trying to control an adult: “How well will this work?”
When his interviewer pushed back on the metaphor — suggesting the adult should guide the children, not the other way around — Korinek turned the question back on them: “So you’re in favor of AI taking over in a few years?”
The Welt journalist responds: “I’m just trying to think your metaphor through further. The adult should please take the children by the hand, not the other way around. Unless the adult is a psychopath.”
The researcher then responded: “This is precisely what the dispute between utopians and apocalypticists revolves around. Some believe that the adult will be kind to human beings; others think his values may be radically different from ours. The children in kindergarten are not always happy. Sometimes they don’t like what the adult tells them, even though it’s for their own good. We may soon find ourselves in this situation.”
