If you are 23 and looking for your first office job, this research is about you. Stanford economists studied payroll records from ADP, which cover millions of U.S. workers through June 20261. They tracked workers aged 22 to 25 in AI-exposed jobs.
Those are jobs where AI tools can already handle many of the tasks. Employment for that group sits 19 percent below where it would be if it had kept pace with peers in other jobs1. Experienced workers in the same jobs kept up.
Other economists read 2026 and see a steady job market57. So who is right? Each camp has real numbers. They count different people, and that difference explains most of the fight.
Fewer first jobs, same paychecks
The gap comes from hiring. Employers brought in fewer young workers and kept the ones already on staff1. In the paper's words, the shift "operates primarily through reduced hiring of young workers rather than increased separations".
A separation is any exit, from a layoff to a quit. Pay held steady, too1.
Why does that matter? Layoff reports count people who lose jobs. Payroll data also counts the people who join, so it shows the drop when fewer young workers start.
The kind of job matters too. Losses cluster in work where AI does a task in place of a person1. Where AI helps people do their work, employment held flat or rose, most of all for experienced staff.
Erik Brynjolfsson, Bharat Chandar and Ruyu Chen hedge their result with care. The pattern weakens once they control for education1. Part of the split also began before generative AI arrived. "We interpret these facts as early, descriptive indicators—canaries in the coal mine—rather than causal estimates," the three wrote.
Layoff counts tell a louder story
Challenger, Gray & Christmas tracks a different signal. The firm logs the reason each employer gives when it announces cuts2. In July 2026, AI led that list for the fifth month in a row, with 10,970 cuts. That was a third of the month's total.
Andy Challenger, the firm's chief revenue officer, summed up the year: "Tech remains the center of gravity for this year's cuts, and AI is still the reason companies give"2. Visa offered a clean case. It announced a 7 percent cut, tied to an efficiency push in which AI would reshape work. Challenger filed those cuts under AI.
Read the next line from the same report, though. "Naming AI in a layoff announcement can win over investors while pushing current and prospective employees away. That's why the messaging has swung from hedging to aggressively citing it," Challenger said2. A stated reason is partly a pitch to shareholders.
Did the streak last? It ended by September, when AI slid to fifth among cited reasons, at 3,961 cuts3. For the year it still leads, with 120,136 cuts, about 21 percent of the total.
The case that the job market is fine
Torsten Slok is chief economist at Apollo Global Management. On June 9, he published a note titled "Where Is the AI Jobs Crisis?" and answered it himself7. "If AI were triggering a jobs crisis, we would expect job openings to collapse and unemployment to climb, yet the opposite is happening," Slok wrote. By his count, job openings per unemployed worker were back above 1.0 that spring.
Yale's Budget Lab reached a similar view in its Oct. 14 update5. "The occupational mix is not yet changing in ways that clearly align with the introduction of AI into the workforce," the lab wrote. It also found that "Measures of AI usage show no connection to changes in employment or unemployment".
Nathan Goldschlag of the Economic Innovation Group adds a caution for both camps: the data itself is thin6. EIG put the stakes this way in its July 2 introduction to his essay: "Partial answers won't do — not at a time when policymakers are struggling to catch up with the sweeping consequences of AI for workers and businesses. Without improved measurement, they risk getting the policy response wrong".
Where the two sides meet
Ask a narrower question, and most of the fight goes away. Is AI cutting jobs across the whole economy? Yale and Apollo find little sign of that yet57.
Is it cutting first jobs for young people in exposed work? Stanford says yes1.
The New York Fed's survey backs both answers. Published Sept. 1, it found 61 percent of service firms in New York and northern New Jersey using AI4. That share was 40 percent in 2025 and 25 percent in 2024.
Layoffs stayed rare: 4 percent of service firms reported AI layoffs, up from 1 percent a year earlier4. Hiring is where firms pulled back. About 15 percent said AI led them to hire fewer workers than planned.
15% of service firms hired fewer workers and 4% reported AI layoffs, as AI use reached 61%
Share of service firms in the New York Fed's survey published Sept. 1, 2026, by answer, percent
Each bar is the share of service firms giving one answer: the gray bar is firms using AI, the magenta bar is firms that hired fewer workers, and the blue bar is firms reporting AI layoffs.
Source: Liberty Street Economics (Federal Reserve Bank of New York) [4]. Chart by Ryan Elliott Dennis.
The numbers behind this chart
| Item | Value | Note |
|---|---|---|
| Use AI | 61% | Up from 40% in 2025 and 25% in 2024 |
| Hired fewer workers than planned | 15% | Brought on fewer workers than they otherwise would have |
| Reported layoffs tied to AI | 4% | Up from 1% a year earlier |
The bank's economists drew the main lesson in one sentence: "Retraining employees in response to AI remains the primary way firms are adjusting their workforces"4. A second line in the same post points at Stanford's work. "one recent study suggests entry-level workers may be affected significantly, as AI can substitute for routine tasks often performed by newer employees, potentially creating barriers to workforce entry even as it enhances productivity for experienced workers," they wrote.
What to do with this
If you hire, count your entry-level hires for 2024, 2025 and 2026, team by team. That count is your own payroll data, and it answers the question for your company before any national survey can.
Starting out? Ask how a team uses AI before you join it. Stanford found employment falling where AI does the task and holding up where AI helps the person1. In an interview, ask which tasks the team has already handed to AI tools.
By the numbers
- 19 percent: the employment gap for AI-exposed workers aged 22 to 25, in Stanford's Aug. 12, 2026 revision1.
- 10,970: AI-attributed job cuts Challenger counted in July 2026, a third of that month's total2.
- 120,136: cuts tied to AI through September 2026, about 21 percent of the year's total3.
- Share of service firms using AI in the New York Fed's survey published Sept. 1, 2026: 61 percent4.
- Service firms reporting AI layoffs in that survey: 4 percent, up from 1 percent a year earlier4.
What to watch
Three trackers will test the gap in public. Stanford's team now publishes AI Economic Indicators for ongoing tracking1. Yale's Budget Lab plans to update its analysis regularly5.
The New York Fed has asked firms about AI each August since 20244. In that survey, watch the 15 percent of service firms hiring fewer workers against the 4 percent reporting layoffs. A rise in the first number is the entry-level story moving into the wider economy.
Correction
AI Lately corrected this piece on Oct. 17, 2026. An earlier version called the Stanford gap 19 percentage points and pinned it to a single 24-year-old. Stanford's paper reports employment 19 percent below its expected pace for workers aged 22 to 25. That version also said the gap widened every month, a claim the paper leaves out. It misnamed two New York Fed authors, Natalia Emanuel and Nick Montalbano. The Fed's 17 percent median share of AI users appeared there as daily use, a term the survey skips. Commas had replaced the dashes in a Stanford quotation, and the dashes are back. Three claims about August layoffs cited a Challenger link to the firm's blog index, and those claims are gone. Yale's findings now come from the lab's Oct. 14 update.
Sources
- Erik Brynjolfsson, Bharat Chandar and Ruyu Chen, "Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence," Stanford Digital Economy Lab, Aug. 12, 2026, https://digitaleconomy.stanford.edu/publications/canaries-in-the-coal-mine/
- Challenger, Gray & Christmas Staff, "Challenger Report: Layoffs Fall, Hiring Picks Up; AI Leads For Fifth Straight Month," Challenger, Gray & Christmas, Aug. 6, 2026, https://www.challengergray.com/blog/challenger-report-layoffs-fall-hiring-picks-up-ai-leads-for-fifth-straight-month/
- Challenger, Gray & Christmas Staff, "Job Cuts Fall in September; Hiring Plans Up 3% Over 2025 On Weak Early Seasonal Hiring," Challenger, Gray & Christmas, Oct. 1, 2026, https://www.challengergray.com/blog/job-cuts-fall-in-september-hiring-plans-up-3-over-2025-on-weak-early-seasonal-hiring/
- Jaison R. Abel, Richard Deitz, Natalia Emanuel and Nick Montalbano, "Businesses Are Using AI to Transform Work, Not Cut Jobs," Liberty Street Economics, Federal Reserve Bank of New York, Sept. 1, 2026, https://libertystreeteconomics.newyorkfed.org/2026/09/businesses-are-using-ai-to-transform-work-not-cut-jobs/
- The Budget Lab Staff, "Tracking the Impact of AI on the Labor Market," The Budget Lab at Yale, Oct. 14, 2026, https://budgetlab.yale.edu/research/tracking-impact-ai-labor-market
- Nathan Goldschlag, "Measuring the Economic Effects of AI," Economic Innovation Group, July 2, 2026, https://eig.org/measuring-ai/
- Torsten Slok, "Where Is the AI Jobs Crisis?," Apollo Global Management, June 9, 2026, https://www.apollo.com/wealth/the-daily-spark/where-is-the-ai-jobs-crisis