Researchers at Yale University say that despite fears about AI taking people’s jobs, there is little evidence that it is actually happening.
Economists at Yale University’s Budget Institute, a nonpartisan policy research group, examined how U.S. employment has changed since the debut of ChatGPT in November 2022 and the subsequent release of other generative AI models.
They didn’t have to worry about anything.
“Overall, the results show that the broader labor market has not experienced any measurable disruption since the release of ChatGPT 33 months ago, allaying concerns that AI automation is currently eroding demand for cognitive labor across the economy,” Martha Gimbel, Molly Kinder, Joshua Kendall and Maddie Lee said in the report’s summary.
AI company leaders are using effective methods to stoke these fears and meet with lawmakers. In May, Anthropic CEO Dario Amodei expressed concern that AI could cut entry-level white-collar employment in half within five years. OpenAI CEO Sam Altman has made similar statements.
And major companies that are cutting jobs, like IBM and Salesforce, are holding themselves out as examples, even though their weeding of employees may be more focused on outsourcing than automation. Small businesses like Fiverr are also citing AI in their layoffs.
Microsoft, an AI cheerleader, recently added fuel to the fire by releasing a report on the jobs most likely to be affected by AI, but later distanced itself from the controversy, saying, “Our research does not conclude that any jobs will be eliminated.”
In Microsoft’s own case, the shakeout is seen as a way to cut expenses and appease investors after massive capital spending on data centers to accelerate its AI ambitions.
There is precedent for the Nothingburger results from the Yale University researchers. In 2023, a United Nations International Labor Organization (ILO) study concluded that generative AI will probably not replace most workers.
A survey of Danish workers published in April found that generative AI had no significant impact on wages or employment. Another similar study published in February found that “the overall employment impact will be modest as the reduction in demand in at-risk occupations will be offset by productivity-driven increases in labor demand in AI-adopted companies.”
There is some conflicting data. A recent study from the Stanford Digital Economy Lab found that new graduates in occupations most exposed to AI experienced a 13% relative decline in employment compared to occupations more insulated from AI.
But so far, the consensus seems to be that generative AI has not had a meaningful impact on the labor market. One reason may be that companies are skeptical about the technology. ®
