Banque de France Study Finds Young People’s Permanent Hiring Falling in AI-Exposed Jobs
Permanent hiring of young workers in France has fallen furthest in the occupations where generative AI can most easily do or assist the work, according to a study by the Banque de France and the Laboratory for Innovation and the Future of Work (LIFT) at HEC Paris, published on October 8. In March 2026, open-ended contract (CDI) hires of 21- to 29-year-olds in the most exposed fifth of occupations were about 28% below their November 2022 level, against 8% in the least exposed fifth, a gap of around 20 points. Franceinfo rounded the drop to nearly 30%.
The authors, Marco G. Palladino, Antonin Bergeaud, Antoine Bertheau, Simon Bunel and Oscar Degos, built an exposure index from France Travail’s directory of occupations, which describes more than 1,500 jobs. A language model judged how far AI could perform or assist each skill, allowing for tacit expertise, physical presence and human contact. The index measures potential rather than actual use, but it correlates closely with usage data from Google’s AI & Economy Atlas. The researchers then matched it with administrative records of hires and departures across four age groups.
The gap narrows for 30- to 39-year-olds and 40- to 49-year-olds and almost vanishes from age 50. Contract endings fell too, though by less: for 21- to 29-year-olds they were about 23% below the November 2022 level in the most exposed occupations, against 6% in the least exposed. Fewer arrivals and fewer departures mean slower turnover, with young people bearing most of it.
What it means in practice
The adjustment shows up in who gets hired rather than in headcount. In a sample of about 6,000 firms, of which nearly 600 reported intensive AI use in early 2026, young permanent hiring fared worse at heavy users, yet their total employment did not diverge clearly from other firms. In France, where ending a CDI is more tightly regulated than in many countries, slowing recruitment or not replacing leavers is an easier lever than layoffs.
The authors are careful about causation. Monetary tightening, the post-Covid correction in tech and remote work hit many of the same jobs from 2022, and firms that adopt AI may differ in other ways. Studies abroad are mixed: research in the United States finds a similar hiring squeeze, while work in Norway and Denmark finds no clear average effect. The bigger concern is long term, since fewer entry-level roles could mean less learning on the job. That puts weight on training efforts such as the €80 million Impulsion IA workplace skills plan and on how small firms adopt drafting assistants.
What to watch: whether the trend holds in the next data releases, whether it spreads to older workers, and whether employers and France Travail adapt apprenticeship and junior hiring schemes.