This Nobel laureate sees AI replacing only 5% of human work in next 10 years

AI may replace only about 5% of the work currently done by humans over the next 10 years, according to Nobel laureate Daron Acemoglu, who said the technology should be developed to complement workers rather than pursue full-scale automation.

Acemoglu, in an article published by The Humanist Review, said his estimate was based on the capabilities of AI models and the pace at which new technologies are typically adopted.

"Two years ago, I estimated that within 10 years, we shouldn’t expect much more than about 5% of what humans do to be replaced by AI," he said.

Microsoft AI CEO Mustafa Suleyman shared the article on X as he announced the launch of The Humanist Review, a new magazine exploring the future of AI.

Acemoglu said he had estimated that by the end of the 2020s, AI would be capable of automating simple office and cognitive tasks accounting for about 20% of work in the US economy.

However, only about one-fourth of those tasks that could be fully automated were expected to actually be automated within 10 years because of the time needed for technological diffusion.

Acemoglu said his earlier forecast had also envisaged GDP increasing by about 1.5% over 10 years.

"Since then, advances in models have been more rapid than many, including me, predicted," he said.

The economist said AI's impact on jobs and productivity would take time because achieving the reliability and versatility needed to automate entire occupations remained difficult.

"AI’s effects will come only slowly, because achieving the kind of reliability and versatility that is required for automating whole jobs isn’t here yet," he said.

He also said evidence from companies using AI in real-world settings was less impressive than the performance of AI systems in controlled environments.

"Several studies and reports have failed to find much in the way of productivity gains from adoption of AI by most companies," Acemoglu said.

He pointed to the continued need for professionals such as radiologists, journalists, paralegals, accountants and office workers despite earlier predictions that AI would rapidly eliminate such occupations.

“We are making the mistake of attempting to force distinctly different artificial intelligence to mimic human intelligence,” the economist said.

"When two things are different, the natural way to proceed is to complement one with the other—not to pretend those differences don’t exist and strive for having one take over everything," he said.

Acemoglu has proposed "pro-worker AI", under which AI would make people better at their jobs, expand their capabilities and create new skills rather than replace them.

"This is the basis of what I have called 'pro-worker AI': rather than pursuing full-scale automation, build AI that makes humans better at their jobs and creates new skills and expertise for them," he said.

Acemoglu said AI could, for example, help teachers identify where individual students were struggling and reorganise classrooms to provide more targeted support, rather than attempting to automate teaching altogether.

"If, instead of running after AGI dreams, we prioritize helping workers acquire better information for better problem-solving across different tasks and create new tasks and capabilities for humans, AI could have much more positive outcomes," he said.

He said the industry's current focus on automation was driven by business models, concentration in the technology sector, investment incentives and an ideological preoccupation with artificial general intelligence.

"The most important tool for steering AI onto a better path is not a silver bullet policy: there is none. It is a change of perspective," he said.