By TOISTHE| Published: October 26, 2026 | 12 min read
Across surveys, job boards and campus conversations, a growing number of students say they fear AI is about to take their future. But is that fear backed by the numbers? Students are worried because entry-level hiring really has weakened, and young workers in AI-exposed jobs are the clearest casualty so far. At the same time, the best evidence says AI is only part of the story. This guide goes through the data, the competing explanations, and what the evidence suggests students can actually do.
Is AI killing entry-level jobs? Partly. Stanford research finds employment among 22-to-25-year-olds in the most AI-exposed jobs is about 19% below where it would be otherwise, mostly because companies are hiring fewer young workers. But the New York Fed and Yale's Budget Lab say AI is not the main driver of the wider hiring slowdown, and find no clear economy-wide AI disruption yet.
Key takeaways
- The pain is real: About 5.6% of recent US college graduates are unemployed and 42% are underemployed (New York Fed, Q2 2026).
- The sharpest AI signal is for young workers: Employment of 22-to-25-year-olds in highly AI-exposed jobs is about 19% below trend (Stanford, August 2026).
- The mechanism is hiring, not firing: Companies are mostly not laying off existing staff. They are hiring fewer juniors.
- AI is not the only suspect: The New York Fed attributes about 64% of the rise in young-graduate unemployment to remote work.
- Students are bracing for it: Roughly two-thirds are pessimistic about the job market, and more than half say AI makes them feel worse about it.
1. Why are students so worried?
Start with the surveys, because the mood itself is part of the story.
- CNBC and SurveyMonkey (May 2026): Two-thirds of students are pessimistic about the job market, and 56% say AI makes them more pessimistic. 62% say AI has made college feel less worthwhile.
- Gallup, Walton Family Foundation and GSV Ventures (April 2026): 31% of Gen Z say AI makes them feel angry, up from 22% a year earlier, while excitement fell from 36% to 22%. Among employed Gen Z respondents, 48% say AI's risks outweigh its benefits at work, up 11 points.
- Handshake (Class of 2026 survey): More than 60% of seniors were pessimistic about their careers. Only 24% of rising seniors believed generative AI would create jobs, compared with more than half of hiring managers.
- Majors are changing: A separate Gallup report found about half of college students have considered changing majors because of AI, and roughly one in six have already done so.
Students are also reading a harder market, not only a scarier headline. Handshake's data showed job postings on its platform down 16% year over year and applications per job up 26% at the time of its survey.
2. What do the hard numbers say?
| Measure | Latest figure | Source |
|---|---|---|
| Unemployment, recent grads (22-27) | About 5.6% | New York Fed, Q2 2026 |
| Underemployment, recent grads | 42% | New York Fed, Q2 2026 |
| Recent-grad unemployment in March 2019 | 3.6% | New York Fed |
| Employment gap, ages 22-25, AI-exposed jobs | About 19% below trend | Stanford Digital Economy Lab, Aug 2026 |
| Overall US unemployment | Roughly 4.2% (September) | US Bureau of Labor Statistics, as reported |
The Stanford "canaries" study
The most cited research is "Canaries in the Coal Mine?" from Stanford's Digital Economy Lab, using payroll data from ADP. Its August 2026 revision found that employment among workers aged 22 to 25 in highly AI-exposed occupations is about 19% below where it would be had it kept pace with similarly aged workers in less-exposed jobs. A year earlier that gap was 13% on data through July 2025, and it has widened with each revision. Experienced workers show no comparable gap.
Two details matter. First, the adjustment appears to work mainly through reduced hiring of young workers rather than more layoffs, so people already in these jobs are mostly keeping them. Second, the declines are concentrated where AI mostly substitutes for human work, with customer service and software development named as examples. Where AI mostly complements workers, employment is flat or rising.
What government data suggests about graduates
Research using US Census data, reported by Fortune in September, followed graduates in the most AI-exposed decile of majors. Their odds of holding a job one quarter after graduation fell by five percentage points after ChatGPT's release, and their initial earnings dropped 13%. One of the authors said the decline in hiring, not job loss, is the primary cause of the unemployment. Separate analysis by the Strada Institute and Burning Glass Institute found 52% of graduates were working in jobs that do not require a degree within a year of graduating, and 45% were still there a decade later.
3. Is AI really the cause?
This is where honest reporting has to slow down. Several serious sources point the other way.
| Evidence AI is hitting young workers | Evidence it is not the main cause |
|---|---|
| Stanford finds a 19% relative employment gap for ages 22-25 in AI-exposed jobs, widening over time, concentrated where AI substitutes for tasks. | The New York Fed found a broad hiring slowdown not concentrated in AI-exposed entry-level jobs, and wrote that AI is not the main driver of it. |
| Census-based research shows worse early outcomes for graduates in the most AI-exposed majors since ChatGPT's release. | New York Fed economists estimate remote work accounts for about 64% of the rise in young-graduate unemployment, because employers find it harder to train juniors on distributed teams. |
| A 41-country Stanford working paper reported that multinational affiliates adopting generative AI grew senior employment by 6.7% while the junior share fell by about two percentage points. | Yale's Budget Lab (updated September 2026) finds no clear sign of AI-related disruption across the labor market as a whole. |
Stanford's researchers have also addressed one common alternative, interest rates. In a February 2026 update they said existing evidence does not suggest rates explain the disproportionate fall in entry-level hiring in AI-exposed jobs. A career expert quoted by Scripps News offered a middle view: many organizations are not removing jobs because of AI, but are "hitting the brakes" on hiring.
The fair reading: there is a genuine hiring slowdown for young graduates, AI is probably contributing in specific occupations, and the share of the blame that belongs to AI is still disputed. Anyone claiming a certain answer in either direction is ahead of the evidence.
4. Which jobs are most exposed?
Stanford's analysis points to work that relies on codified knowledge: formal, standardized, documented knowledge that can be taught from textbooks or written procedures. That is exactly what AI systems learn well. Experienced workers in the same occupations rely more on tacit knowledge built through judgment and experience, and show no comparable decline.
Examples named in the research include software development, customer service and clerical work. In the UK discussion, analysts also mention data entry and document review. The pattern worth remembering: the more a job is made of standard, learnable tasks, the more exposed its entry-level version appears to be.
PwC's 2026 barometer, as cited in one analysis, found AI-exposed entry-level roles are seven times more likely to ask for senior-level skills. Whether that is a lasting shift or a temporary squeeze is not yet clear.
5. How are employers responding?
Employer surveys paint a more nuanced picture than the student mood suggests.
- An Adzuna survey of 750 HR leaders across the UK, US and India found businesses believe AI now carries out roughly a third of entry-level tasks, rising to 39% in the UK.
- 85% of those leaders still consider entry-level roles essential, and 96% expect them to evolve into "AI supervision" positions over five years rather than disappear.
- 94% expect AI to create entirely new entry-level roles that do not exist today.
- At a techUK panel, IBM said it is tripling its US entry-level hiring because of AI, viewing early-career hires as well placed to work with new technology. Speakers also warned that cutting junior pipelines now risks a shortage of experienced managers in five to ten years.
- In one analysis, US employers named AI in about 120,000 announced job cuts from January to September 2026, roughly a fifth of the total, though that does not show AI caused them.
6. Is it only a US problem?
No. The UK data shows a similar squeeze, with a twist in the measurements.
- Adzuna recorded 8,383 UK graduate vacancies in July 2026, down 45.6% from 15,397 a year earlier, the lowest in a decade according to reports.
- The Institute of Student Employers, which surveys employers, reported a smaller 8% dip in graduate hiring and growth in apprenticeships.
- The Work Foundation at Lancaster University found more than a third of UK employers reduced entry-level opportunities for younger workers over the past year.
- Economic pressures, hiring freezes and slower growth are also cited alongside AI.
The gap between job-board data (a big fall) and employer surveys (a smaller fall) is a reminder that different methods give different answers.
7. What can students actually do?
None of this is a guarantee, and nobody can promise a safe path. But the research suggests several practical directions.
- Build what is hard to codify. Judgment, communication, client work and domain context are the "tacit knowledge" that experienced workers show and AI struggles to replace. Internships, projects and real work help build it.
- Be a confident AI user. Employers increasingly expect juniors to work with AI tools. Handshake found most students use AI regularly, but confidence and skill vary widely.
- Prefer roles where AI complements. Stanford found employment flat or rising where AI mainly assists workers. When comparing options, ask whether the work leans on AI as a helper or as a replacement.
- Keep your options broad. Some students are moving toward careers seen as more automation-resistant, and apprenticeships are growing in the UK.
- Look beyond the big graduate schemes. At the techUK panel, speakers noted most jobs do not come through large graduate programmes.
- Show proof of skill. Employers are moving toward skills-based hiring, so portfolios and work samples can matter alongside a degree.
Note: This is general information, not personal career advice. For decisions about your degree or career, consider talking to a careers adviser.
8. What to watch next
- November: The New York Fed updates its recent-graduate unemployment and underemployment data with Q3 2026 figures.
- Monthly: Yale's Budget Lab continues to update its AI labor-market tracker.
- Stanford: Further revisions of "Canaries in the Coal Mine?" as more payroll data arrives.
- Hiring season: Next spring's graduate recruiting numbers from employers and job boards.
9. Frequently asked questions
Is AI taking entry-level jobs?
In some occupations, yes. Stanford finds employment for 22-to-25-year-olds in highly AI-exposed jobs is about 19% below trend, driven by reduced hiring. But the New York Fed and Yale's Budget Lab say AI is not the main driver of the broader slowdown.
What is the unemployment rate for recent college graduates?
About 5.6% for ages 22 to 27 in the second quarter of 2026, according to the New York Fed, up from 3.6% in March 2019. Underemployment was 42%.
Which jobs are most at risk from AI for new graduates?
Research points to entry-level work built on standardized, codified tasks, such as parts of software development, customer service and clerical work.
Is a college degree still worth it?
Graduates have historically had lower unemployment and higher lifetime earnings than non-graduates, but 62% of students in one survey say AI has made college feel less worthwhile. The answer depends on the field, cost and individual goals.
Are students right to be worried?
The hiring squeeze is real, so the concern is grounded. How much of it is caused by AI, as opposed to remote work, the economy and employer caution, is still debated.
What skills help against AI job displacement?
Skills that rely on judgment, communication and domain experience, plus the ability to use AI tools well, are repeatedly named by researchers and employers.
The bottom line
The honest summary is neither "AI is killing the future" nor "nothing to see here." Young graduates face a weaker market than their older siblings did, the sharpest AI-linked decline is among 22-to-25-year-olds in the most exposed jobs, and the mechanism is fewer new hires, not mass layoffs. At the same time, remote work, a cautious economy and employer behavior explain a large part of the picture. The next few quarters of data, especially the New York Fed's November update, will show whether the gap keeps widening.
Related reading
- Fed Meeting October 28: What to Expect and Why Oil Matters
- Claude AI Bypassed a Georgia Paywall: What Happened and What We Know
Sources: Federal Reserve Bank of New York (college labor market dashboard, Q2 2026; Liberty Street Economics on remote work, June 2026); Stanford Digital Economy Lab, "Canaries in the Coal Mine?" (August 2026 revision) and February 2026 update; Yale Budget Lab AI labor-market tracker (September 2026); Fortune (September 17, 2026) on Census-based research; CNBC and CNBC/SurveyMonkey (May and September 2026); NPR (August 18, 2026); Gallup/Walton Family Foundation/GSV Ventures (April 2026); Handshake Class of 2026 outlook via Inside Higher Ed; Scripps News; Adzuna, Institute of Student Employers, Work Foundation, techUK and Consultancy.uk on UK data. Figures are as reported by the cited sources and may be revised.
0 Comments
please do not enter any spam link in the comment box