Key Takeaways: A growing body of payroll data shows AI is eliminating entry-level jobs even as aggregate employment holds steady, deepening a divide between three prominent economists over whether the labor market faces a structural shift or a temporary adjustment.
A growing body of payroll data shows AI is eliminating entry-level jobs even as aggregate employment holds steady, deepening a divide between three prominent economists over whether the labor market faces a structural shift or a temporary adjustment.
The debate over whether artificial intelligence will destroy jobs has moved from hypothetical to empirical, with fresh payroll data showing a 3.8% annual decline in AI-exposed roles for workers aged 22 to 25 — even as overall employment in those occupations barely budges.
"Whatever it is, it's not going away," Erik Brynjolfsson, a Stanford economist who leads the Digital Economy Lab, told Fortune, referring to the persistent decline in early-career hiring since ChatGPT's launch.
The data, drawn from ADP's payroll records covering roughly one in six American workers, shows the effect has actually intensified over time. Employment for young workers in highly AI-exposed occupations fell 2.8% year-over-year through April 2024, then accelerated to more than 4% annually through April 2026. By contrast, the least-exposed jobs for the same age group are growing at 2% annually. Mid-career workers ages 31 to 34 are also contracting, down 1.7% year-over-year, while workers ages 35 to 40 are growing at 2%.
The divergence has fueled a public rift among top economists. Brynjolfsson sees the Industrial Revolution for the mind unfolding 10 times faster than its predecessor. MIT Nobel laureate Daron Acemoglu calls much of the AI productivity discourse "brainless" — speculative to the point of fiction. Nvidia CEO Jensen Huang, meanwhile, dismisses job-loss fears as "complete nonsense," arguing AI will make companies more productive without reducing headcount. The stakes are enormous: if Brynjolfsson is right, millions of young workers could face a permanently damaged career on-ramp.
The disagreement is not about whether AI is affecting work — it's about the magnitude and timeline. Brynjolfsson's Canaries Dashboard, built in partnership with ADP Research and updated through April 2026, now covers 4.6 million workers across more than 730 occupations. The dashboard processes payroll data in near-real time, offering what Brynjolfsson calls "timely, trusted evidence" for a debate that has been dominated by speculation.
The mechanism behind the early-career decline is straightforward. AI absorbs tasks before it absorbs jobs, and the tasks it reaches first are the ones that require the least experience: retrieving, summarizing, scheduling, formatting. These are disproportionately the tasks assigned to junior employees. Senior workers have accumulated hard-to-codify, job-specific skills that still buffer against displacement.
ADP chief economist Nela Richardson, Brynjolfsson's research partner, has argued that the distinction between automation and augmentation is the key variable. Occupations where AI augments human work show more enduring employment growth; those where AI automates tasks outright show contraction. "In the aggregate, AI's impact on jobs remains modest," Richardson wrote in a June 16 blog post. But when measured by career stage, she continued, "dramatic differences emerge."
Brynjolfsson has stress-tested the finding against every major counter-argument. The interest rate hypothesis points the wrong direction — the most rate-sensitive occupations, like construction, have the lowest AI exposure. He removed the entire tech sector. He isolated remote-work effects. The pattern held every time.
The Productivity Bet
Brynjolfsson has a 10-year wager with Northwestern economist Robert Gordon on longbets.com that productivity will be significantly higher by the end of the decade. "I'm already ahead," Brynjolfsson said. "And I always figured it was backloaded because of my J-curve theory."
Acemoglu, for his part, has not softened his skepticism. He recently told Fortune that much of the AI productivity discourse is speculative to the point of fiction. Yet both economists agree on one point: AI should be deployed to complement human workers rather than replace them. The question is whether market incentives will push in that direction.
For investors, the debate carries direct portfolio implications. Companies that deploy AI to automate entry-level functions may see near-term margin expansion — but at the cost of a depleted talent pipeline. Nvidia, trading at roughly 35x forward earnings, has the most to gain if AI adoption accelerates regardless of labor market outcomes. Firms in HR technology and workforce training, such as Upwork or Coursera, could benefit if reskilling becomes a corporate priority. Conversely, businesses heavily reliant on junior white-collar labor — consulting, legal services, accounting — face the most structural risk.
The data has not yet moved markets. But as Brynjolfsson's dashboard updates monthly with fresh payroll figures, the debate is shifting from "if" to "how much" — and that shift alone is a signal for investors to pay attention.
This article is for informational purposes only and does not constitute investment advice.
