The first place AI shows up in the job market is likely to be the hiring of young people into jobs AI can partly do. This page brings together the main evidence on the AI impact on entry-level jobs, US and UK, with dates, sources and each study’s own caveats. It includes studies that find little or no effect.
Where the evidence stands: The strongest evidence of harm is narrow: several US datasets (Stanford/ADP payroll, the Dallas Fed, Revelio, Anthropic's survey work) find that fewer people in their early 20s are being hired into the jobs AI can do most of, with Stanford putting the gap at 19% by mid-2026. None of them finds broad job loss or a rise in unemployment among exposed workers, and the effect works through slower hiring of newcomers, not layoffs of people already in work. Other researchers (Yale Budget Lab, the Economic Innovation Group, Frank and colleagues) argue the slide began before ChatGPT or tracks interest rates and the tech-hiring cycle, and every study's authors say their results are not proof that AI caused them. In the UK, graduate postings and youth unemployment are weak, but the official figures do not separate AI from other causes, and one study (King's College London) links AI exposure to fewer junior roles.
Studies that find an effect
Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence (August 2026 update)
Employment of 22-25-year-olds in the most AI-exposed jobs is about 19% below where it would be had it kept pace with less-exposed peers, up from a 15% gap in July 2025 data; experienced workers show no comparable gap. The shortfall comes mainly through less hiring of young people rather than more separations, and it is concentrated in jobs where AI is used to automate tasks rather than help workers.
SourceYoung workers' employment drops in occupations with high AI exposure
Employment of 20-24-year-olds in the most AI-exposed occupations has fallen about 13% since November 2022, mainly because fewer young people are moving into those jobs from outside the labor force, not because of layoffs. Even if all of that decline became unemployment, it would add only about 0.1 percentage point to the national rate.
SourceAI Labor Market Tracker, August 2026
Employment in the most AI-exposed occupations is about 6% lower than in the least exposed since October 2022, and for workers aged 22-25 the gap is 19%, against about 5% for older workers. Demand has fallen more for junior roles than senior roles in exposed occupations.
SourceNew study reveals early impact of AI on job market in UK
UK firms most exposed to AI cut total employment by about 4.5% on average, with the drop concentrated in junior positions (down 5.8%), while average pay at those firms rose.
SourceMixed findings
Labor market impacts of AI: A new measure and early evidence
There has been no systematic rise in unemployment for workers in the most AI-exposed jobs since late 2022. For 22-25-year-olds, the rate of starting a job in exposed occupations fell about 14% compared with 2022, a result the authors call just barely statistically significant; there was no such drop for workers over 25.
SourceThe Labor Market for Recent College Graduates
Conditions stayed difficult for recent graduates through the second quarter of 2026: their unemployment rate was about 5.6%, above the rate for all workers, and 42% were underemployed. Computer engineering and computer science graduates had among the higher unemployment rates by major in 2024.
SourceIs AI responsible for the rise in entry-level unemployment?
Entry-level postings fell about 35% below January 2023 levels, with highly AI-exposed entry-level roles down about 40% against 33% for low-exposure ones. More AI exposure was linked to lower demand for juniors but higher demand for seniors.
SourceNo Country for Young Grads
Unemployment among 20-24-year-olds with at least a bachelor's degree rose from 5.2% in 2018-19 to 6.2% in the two years to June 2025, and 52% of the Class of 2023 were in jobs not needing a degree a year after graduating. The report treats AI as an accelerant of leaner post-pandemic staffing and a growing supply of graduates, not the sole cause.
SourceEducated but unemployed, a rising reality for US college grads
Higher unemployment among recent graduates is mainly a structural shift in tech hiring alongside strong growth in the supply of graduates, with signs that AI is displacing some entry-level positions.
SourceAI and Job Postings: From Destruction to Creation?
US software development postings have risen almost 15% since late February 2025 while overall postings fell 7%, but the rebound is tilted to experience: 71% of the increase between May 2025 and May 2026 came from senior roles.
Source2026 Global AI Jobs Barometer
In US job ads for entry-level roles (0-2 years' experience), the most AI-exposed roles were seven times as likely as the least exposed to ask for skills once reserved for senior staff, such as leadership and strategic decision-making. Within the most AI-exposed quarter of jobs, postings for entry-level roles that added these skills rose 35% from 2019 to 2025, while postings for the rest fell 10%.
SourceYouth unemployment statistics
Youth unemployment in the UK was 16.2% in April to June 2026, up from 14.3% a year earlier and close to its highest since 2014, with 739,000 young people unemployed. The library's web page, updated for May to July 2026, puts the number at 751,000, up 109,000 on the year.
SourceIndeed's 2026 Mid-Year UK Jobs & Hiring Trends Report
UK graduate postings were about 7% lower than a year earlier as of 10 July 2026 and at their lowest for the time of year since 2020, while a record 9.4% of UK postings mentioned AI. Indeed puts the weakness down mainly to employer cost pressures and uncertainty.
SourceEntry-level work reshaped, not replaced (Student Development Survey 2026)
Most UK graduate employers (87%) expect AI to change graduate and apprentice roles, but 40% expect to replace none of them with AI over three years and 42% only 1-10%; 18% expect to replace more than 10%.
SourceStudies that find little or no effect
Tracking the Impact of AI on the Labor Market
With data through August 2026, Yale finds no clear evidence of labor-market disruption from AI: occupational churn, AI exposure among the unemployed and usage measures are flat, within historical ranges, or on pre-AI trends. In its January 2026 update, the gap between the jobs held by recent graduates and older graduates stayed mostly within a 30-33% band since 2021.
SourceAI and Jobs: The Final Word (Until the Next One)
The authors find no meaningful sign of AI-driven job loss: from 2022 to early 2025 unemployment rose 0.30 percentage point for the most exposed workers and 0.94 point for the least exposed. Unemployment rose for recent graduates in both exposed and unexposed fields.
SourceLooking for the Ladder: Is AI Impacting Entry-Level Jobs?
The authors argue the timing does not fit AI: much of the decline came within months of ChatGPT's release, before firms could plausibly redesign work, and is better explained by the sharpest monetary tightening in four decades.
SourceAI-exposed jobs deteriorated before ChatGPT
Unemployment risk in AI-exposed occupations started rising in early 2022, months before ChatGPT, and graduates from 2021 onward were already entering exposed jobs at lower rates. Graduates whose courses were more AI-exposed got better-paid first jobs and found work faster after ChatGPT.
SourceAI and the UK labour market: the evidence so far
Three years after ChatGPT, the review finds no detectable job losses in AI-exposed UK occupations and entry-level pay in exposed jobs roughly flat relative to other jobs, concluding that AI is not replacing labor at scale at current adoption rates.
SourceWhat the research can’t tell us yet
- Cause. Interest rates, the end of pandemic hiring and the tech-hiring cycle all hit young workers too. The Stanford authors call their figures descriptive indicators, not causal estimates, and note that some of the gaps predate generative AI.
- Permanence. Fewer junior hires now could mean fewer experienced workers later, or employers could change how they train people.
- Your job. Averages across exposed jobs say little about any one employer or role.
What to do about it
How we pick the studies
We include research that measures hiring, employment or job postings for young or early-career workers and looks at exposure to AI, from universities, central banks, government statisticians and companies that publish their method. Each card links to the source. Evidence checked October 3, 2026.