AI Is Erasing the Entry-Level Job, Stanford Data Shows

The clearest evidence yet that AI is reshaping the bottom rung of the job market arrived this week, and it points in one direction: the youngest workers are getting squeezed first. In an analysis published by Fortune, Stanford economist Erik Brynjolfsson reported that employment among workers aged 22 to 25 in the most AI-exposed occupations is now shrinking by 3.8% a year — and that the decline has been accelerating, not fading.

The strength of the finding comes from the data behind it. Rather than relying on surveys, Brynjolfsson's team used a large, high-frequency administrative dataset from ADP, the largest payroll processor in the United States, covering millions of workers across more than 730 occupations — roughly one in six American workers. That granularity lets the researchers separate effects by age and by how exposed a job is to AI, and the contrast is stark: while the youngest workers in exposed roles are losing ground fast, workers aged 35 to 40 in the same fields actually saw employment grow by about 2%.

The pattern has a straightforward mechanism. AI tools absorb the most easily automated work first — information retrieval, summarization, scheduling, formatting and routine drafting — and those are precisely the tasks handed to junior employees. Senior staff hold expertise that is harder to codify and so remain insulated, at least for now. Crucially, the researchers say the entry-level decline is driven not by mass layoffs but by a collapse in hiring: young people entering the workforce simply cannot find the openings that used to exist. Brynjolfsson notes the effect survives a battery of robustness tests — stripping out the tech sector, isolating remote-work patterns, and controlling for interest-rate sensitivity all leave the result intact.

The project's dashboard is named, pointedly, "Canaries in the Coal Mine." The framing is deliberate: early-career employment is an early-warning signal, not the full danger. As the team puts it, the canaries "didn't stop the danger. They just told you the clock was running." Looked at across all ages, AI-exposed occupations contracted just 0.2% over the year — a headline number mild enough to hide the sharp generational split underneath it.

Brynjolfsson, who flagged this trend early and has faced pushback that it was a blip tied to a cooling tech labor market, says the latest data settles the question for him: "It's not going away." For a generation trying to get a first foothold in white-collar work, the study reframes a familiar anxiety as a measurable trend — and hands policymakers, educators and employers a number they can no longer wave off.