The mass adoption of ChatGPT is yet to have a big disruptive impact on US jobs, contradicting claims by chief executives and tech bosses that artificial intelligence is already upending labour markets.
Research from economists at the Yale University Budget Lab and the Brookings Institution think-tank indicates that, since OpenAI launched its popular chatbot in November 2022, generative AI has not had a more dramatic effect on employment than earlier technological breakthroughs.
The research, based on an analysis of official data on the labour market and figures from the tech industry on usage and exposure to AI, also finds little evidence that the tools are putting people out of work.
The study follows widespread concern that generative AI will spark job losses — and even the disappearance of certain types of work — amid a US labour market that has recently weakened.
“Despite how quickly AI technology has progressed, the labour market over the past three years has been a story of continuity over change,” said Molly Kinder, a senior fellow at Brookings who co-authored the research. “We are not in an economy-wide jobs apocalypse right now, it’s mostly stable. That should be a reassuring message to an anxious public.”
Martha Gimbel, the co-author who heads the Yale Budget Lab, said: “The labour market doesn’t feel great, so it feels correct that AI is taking people’s jobs. But we’ve looked at this many, many different ways, and we really cannot find any sign that this is happening.”
While the analysis suggests ChatGPT — one of the most widely-used text-based forms of generative AI — is rapidly changing the mix of occupations on offer to tech workers, it is not shifting the composition of jobs throughout the entire US economy at a much swifter pace than the arrival of computers and the internet.
“AI has, so far, not defied gravity,” said Kinder. “We are in very early days of companies figuring out how to redesign themselves with this technology.”
The research, which was co-authored with the Yale Budget Lab’s Joshua Kendall and Maddie Lee, also finds limited evidence that graduates in the world’s largest economy have struggled to find work because of the rise of generative AI.
The unemployment rate for those with a bachelor’s degree aged 20-24 soared to 9.3 per cent in August, more than double the 4.4 per cent rate recorded in April. But there is little to distinguish the positions taken by those in the 20-24 age range, compared with those on offer from older graduates aged 25-29.
The lack of change in the occupational mix suggests the difficulties of finding work for those just out of college has little to do with technological change.
The findings conflict with assertions from chief executives and tech bosses that generative AI is swiftly reshaping the labour market.
A new report from the British Standards Institution, which surveyed more than 850 business leaders across eight countries and analysed company annual reports, found 39 per cent of those polled said AI had led them to cut entry-level roles. Another 43 per cent said they were expecting further reductions in a year.
Dario Amodei, chief executive of AI start-up Anthropic, in May warned of the mass elimination of roles in fields such as law, consulting and finance, and claimed the technology had the potential to wipe out half of entry-level jobs. In an interview with Axios, he said unemployment would jump to 10-20 per cent in the next five years — up from a current level of 4.3 per cent.
Meanwhile, OpenAI chief Sam Altman has said AI would eliminate some job categories, such as customer service.
Many economists, however, believe claims by CEOs and the tech industry are — at least on an economy-wide level — overblown.
The AI Shift
Sign up to “The AI Shift newsletter”, a rigorous weekly deep-dive into how AI is reshaping the world of work, with John Burn-Murdoch and Sarah O’Connor
“There is a lot of pressure on managers to do something with AI . . . and there is the hype that is contributing to it,” Daron Acemoglu, Nobel Prize-winning economist and professor of economics at the Massachusetts Institute of Technology. “But not many people are doing anything super creative with it yet.”
Acemoglu added that AI companies were incentivised to drum up attention for their technologies to encourage companies to build the computing capacity they need to create bigger AI models.
A recent study from Goldman Sachs Research estimated that AI adoption could displace 6-7 per cent of the US workforce, but said the impact will probably be “transitory”.
The Yale Budget Lab intends to update their data every month.
“It is an open question [whether or not AI will change the labour market] and we should be monitoring this,” said Gimbel. “But let’s not put the cart before the horse here.”