Back to Stories

Is AI Leaving Young Workers Behind? Two New Studies Say Yes



Laura Nurski
September 4, 2025 - 2 min read

Two new studies offer the clearest signal yet that generative AI may be reshaping early careers – before we even notice its full economic footprint.

The first, by Brynjolfsson, Chandar and Chen (2025), uses individual-level payroll data – covering 3.5 to 5 million US workers across tens of thousands of firms – to document a 13% relative decline in employment for 22- to 25-year-olds in the most AI-exposed (the extent to which an individual, job, organization, or system is affected, dependent on, or vulnerable to AI technologies) occupations. Crucially, these declines occur even as overall employment continues to grow. The effects are concentrated in jobs where AI automates tasks (like customer support and software development), and are not observed in roles where AI tends to augment human work.

Source: Brynjolfsson, Chandar and Chen (2025)

A second study by Lichtinger and Hosseini Maasoum (2025) complements this picture. Using résumé and vacancy data from 62 million workers across 285,000 firms, they track the within-firm impact of generative AI adoption (proxied by “AI integrator” job postings). They find that junior hiring slowed sharply from Q1 2023 onwards in adopting firms, while senior employment continued to grow. The effects are strongest for workers with mid-tier educational backgrounds, suggesting a U-shaped pattern by education level.

Source: Lichtinger and Hosseini Maasoum (2025)

Together, the papers suggest a new gradient in technological change: not just skill-biased or routine-biased, but now also “seniority-biased” technological change. If generative AI is shifting demand away from entry-level roles, this could make early labour market entry more fragile – especially in sectors like customer service, and software development. When the first rung of the career ladder starts breaking, are we ready for the knock-on effects?

While both studies focus on the US, their implications could very well be global. Will labour markets and organisations in Europe respond in the same way? Time will tell – replication studies with EU data should be hitting us soon…


Scan the QR code to view this story on your mobile device.


GenAIFuture Of WorkEmployment