US employers have cited AI for 120,136 announced job cuts this year, more than double the total for 2025. Graduate unemployment has barely moved. In the US, the UK, Canada and Australia, the pressure falls on hiring young people into AI-exposed jobs.
American employers cited artificial intelligence for 120,136 announced job cuts in the first nine months of 2026. The figure comes from a commercial tally by the outplacement firm Challenger, Gray & Christmas. That’s about 21% of all announced cuts, and AI remains the leading reason companies give this year. Technology firms announced 165,925 cuts by the end of September, up from 107,878 a year earlier.
For the whole of 2025, Challenger counted 54,836 AI job cuts. Total announced cuts are falling, though: 573,195 so far in 2026, down 39%. The firm logs announcements and the reasons employers give for them. Unemployment is measured separately, by government surveys.
Economists Robert Fairlie and Jane Wu searched those surveys for an AI effect on recent US graduates. Their working paper for the National Bureau of Economic Research came out in September. It finds that “unemployment rates did not spike in summer 2026 relative to summer months in previous years”. The result holds against older graduates and against young people without degrees.
Where the losses show up
ADP payroll records help explain the gap between those two findings. Erik Brynjolfsson and colleagues at Stanford’s Digital Economy Lab studied 22 to 25-year-olds in highly AI-exposed occupations. Their employment now sits about 19% below where it would be had it kept pace with less exposed peers. The gap was 15% in July 2025. The adjustment works “primarily through reduced hiring of young workers rather than increased separations”, the authors write.
Census Bureau researchers found a sharper effect by degree subject. Graduates in the most AI-exposed tenth of majors saw their chance of a first job fall by five percentage points. Their initial earnings dropped 13%. The authors compare that loss to graduating into a large recession. About half of it comes from graduates moving into lower-paid work in restaurants and retail.
In Texas, the Dallas Fed ranks computer science, computer engineering and languages among the most exposed majors. Nursing, education and psychology sit at the other end. Each 10-point rise in the share of a major’s automatable tasks goes with a 1.7-point relative fall in graduates’ employment rates.
Oracle’s headcount fell from 162,000 to 141,000 in the year to May. Its annual filing says AI deployment has “resulted, and may continue to result, in reductions” to its workforce. The Bureau of Labor Statistics projects 153,700 fewer customer service representative jobs by 2034, a fall of 5.5%. It names AI as one cause.
The same pattern abroad, at a smaller scale
UK job vacancies have nearly halved since their 2022 peak. A Bank of England staff analysis published in August sorted occupations by AI exposure. The most exposed third lost 15% of online adverts, against 6% for the least exposed. Customer service adverts fell 23% and administrative adverts 22%.
Haley Schlicht, the author, suggests firms may be moving from a pyramid built on junior staff towards a “diamond”, hollow at the base. She also warns that pandemic over-hiring, remote work and cost pressures make attribution difficult.
The Office for National Statistics found that around 35% of British firms with ten or more staff used AI by June. In late 2023 the figure was about 12%. About half said AI had no effect on headcount. The UK’s graduate squeeze has deeper roots, as RESET reported in Young, Qualified and Struggling.
In Canada, coding-intensive jobs grew about 15% between November 2022 and December 2025. Yet employment of 15 to 29-year-olds in those jobs barely changed, while it rose nearly 30% for 30 to 49-year-olds, Statistics Canada found. Its authors caution that post-pandemic adjustment, demographics and trade tensions with the US could explain part of the pattern.
Employment in Australia’s AI-exposed occupations rose 5.6% from November 2022 to February 2026, the employment department reported. The least exposed grew 9.5%. Software developer jobs grew 25% over the period. Jobs for 20 to 24-year-olds grew slightly faster than for older workers.
Large language models appear to explain little of young people’s job difficulties so far, the OECD’s Employment Outlook judged in July. It points to weaker labour markets and longer-run shifts in skill needs instead.
Where AI is adding work
Euro area firms that use AI intensively are about 4% more likely to take on staff. A European Central Bank blog drew that from its survey of about 5,000 companies. Small firms drive that effect. For large firms, AI has so far made no difference to employment.
The BLS expects AI to “fuel strong job growth among computer and mathematical occupations”. It projects 267,700 more software developer jobs by 2034, a rise of 15.8%. Data scientist jobs are set to grow 33.5%, and information security analyst jobs 28.5%.
PwC studied more than a billion job adverts in 27 countries for a commercial report. Jobs requiring specific AI skills grew 69%, against 9% for the overall market. Workers with those skills earned a 62% wage premium.
In US data, PwC found adverts for AI-exposed entry-level roles had grown 35% since 2019, while other entry-level adverts shrank 10%. Those roles were seven times more likely to ask for skills traditionally expected of senior staff. PwC counts adverts, which the payroll data from Stanford and the Census Bureau suggest young graduates are struggling to convert into jobs.
US construction must hire 349,000 workers in 2026, Associated Builders and Contractors estimates. AI data centres, already the target of new EU energy rules, and chip plants are among the megaprojects driving that demand. “Skilled trades workers, particularly electricians, are a huge constraint at the moment,” Anirban Basu, the association’s chief economist, told DCD.
Openings in construction and AI engineering call for different skills, in different places, from the junior roles being cut. Next summer’s graduate data will test which reading holds. If unemployment stays flat while hiring into exposed fields keeps falling, the cost will be careers that never start. No layoff tally records those.
Subscribe
Sign-up to receive our newsletter
