TL;DR
The International Labour Organization reports that unemployment among 15-to-24-year-olds climbed to 12.4% in 2025, up from 12.3% two years earlier and equal to 67 million people. It attributes the rise to weak growth and thin job creation, but singles out the disappearance of middle-skill work as the structural problem — and estimates that 6.1% of jobs held by under-30s sit in the occupations most exposed to AI.
The entry rung is the problem
The headline movement is small. What sits underneath it is not. The agency’s report identifies clerical work, administration, sales and factory jobs as the shrinking categories — precisely the jobs that have historically absorbed school and university leavers before they specialise.
Its phrasing is unambiguous: fewer middle-skill posts mean “longer job queues and growing unemployment” for young people holding secondary-level qualifications. Where those roles once functioned as an entry rung, candidates now compete for a shrinking number of them.
A parallel measure — under-25s outside work, study or training — also drifted up, to 20%, or 257 million-plus. Deterioration was near-universal, appearing in eight of eleven subregions and across income brackets. Worst affected were the Arab States on 26.2%, with Northern Africa close behind at 22.6%. North America deteriorated fastest, moving from 8.3% to 9.8%. Britain sits inside the northern, southern and western Europe grouping, which registered 15%.
What the AI estimate actually says
The 6.1% figure describes exposure, not loss. Should a tenth of those positions vanish outright, the agency puts roughly 5.6 million young workers at risk of unemployment, forced career changes or leaving the workforce, concentrated in wealthier economies. It concedes the long-run picture is uncertain while cautioning against dismissing the risk.
Set against that, science, engineering, healthcare and IT roles continue growing in most countries — which is a skills-distribution problem rather than a jobs-total one.
Looking forward
For UK readers the useful pairing is with domestic evidence rather than the global aggregate. Our recent coverage found half of British small firms see no role for AI in their business at all, while employers elsewhere are already redirecting AI-freed hours into higher-margin advisory work. Both patterns squeeze the same rung: firms that automate junior tasks stop hiring for them, and firms that ignore AI entirely are not creating the roles either.
The policy question this hands UK ministers is narrow and awkward. Skills funding assumes an entry-level job exists at the end of the training. Where the ILO’s middle-skill decline continues, that assumption needs evidencing rather than asserting.