Artificial intelligence (AI) has fueled fears about the future of work, but a new study suggests the U.S. job market has yet to experience major disruptions.
Researchers at the Yale University Budget Lab and the Brookings Institution found little evidence that generative AI has dramatically altered employment since OpenAI launched ChatGPT in November 2022.
“Despite fears of an imminent AI jobs apocalypse, the overall labor market shows more continuity than immediate collapse,” wrote Molly Kinder, a senior fellow at Brookings who co-authored the research.
The researchers looked at the share of workers in jobs with high, medium and low AI “exposure” and found those levels have remained largely steady since ChatGPT’s launch. They also examined whether AI-displaced workers were showing up in unemployment stats but found no pattern of rising AI exposure among the jobless.
That doesn’t mean AI hasn’t had any impact over the past three years. The authors said their analysis is consistent with emerging evidence suggesting AI may be contributing to unemployment among early-career workers.
Separate research from Stanford University has found that employment among early-career workers (ages 22-25) in the most AI-exposed occupations has dropped 13 percent since the widespread adoption of generative AI.
Tech leaders have also warned about looming changes, some of which have already arrived. Salesforce CEO Marc Benioff recently said AI enabled him to cut thousands of customer support roles this year.
Meanwhile, the CEO of Anthropic told Axios in May he expects AI could wipe out half of all entry-level white-collar jobs — and spike unemployment to 10-20 percent in the next one to five years.
Still, broad AI-driven labor-market turbulence hasn’t materialized yet. At a news conference last month, Federal Reserve Chair Jerome Powell said AI may be having “some effects,” but ventured it’s “not the main thing” driving current employment trends.
The Yale team stressed that it’s too soon to tell how disruptive AI will be for jobs in the long run, noting that previous technological shifts often played out gradually.
“Historically, widespread technological disruption in workplaces tends to occur over decades, rather than months or years,” they wrote.
Long-term projections do point to structural shifts ahead. The Labor Department expects major declines over the next decade, including about 310,000 fewer cashier jobs by 2034 (a 10 percent drop), nearly 180,000 fewer office clerks (-7 percent) and more than 150,000 fewer customer service representatives (-5.5 percent). Not all of those losses will be tied directly to AI, but the technology is likely to accelerate some of the changes already underway.
For now, though, the research suggests that AI’s impact on the U.S. job market has been more muted than many feared.
