Research update · 7 October 2026
Danish chatbot adopters did not show a sharper early-career drop
Young workers lost ground in several AI-exposed occupations. But in a Danish survey linked to administrative records, workplaces encouraging AI chatbots did not have a larger fall in their early-career share than other workplaces through December 2024.
What was compared
Humlum and Vestergaard surveyed about 25,000 Danish workers in 11 AI-exposed occupations and linked responses to monthly employment records. They compared workplaces that encouraged chatbot use with other workplaces. The early-career share moved no differently between the groups; the authors say their estimates rule out differences larger than about one-third of a percentage point. This is an employer-use comparison, not a comparison of people who took an AI course.
The same underlying Danish working paper was already in our evidence map for its earnings and hours results. The new author explanation adds detail to that study. It is not an independent replication.
What it does not settle
Workplace encouragement is not randomized adoption, and the data stop in 2024. The result does not explain why early-career shares fell in the exposed occupations. It does not erase U.S. graduate-entry declines or a 41-country affiliate study that reports a lower junior share after firms post GenAI roles. That study uses a different adoption proxy, sample and end date. A direct comparison needs matched definitions and periods.
There is no personal risk estimate here. The next evidence check is whether measured AI use, junior hires and separations move together for comparable employers over a longer period.