We examine youth employment trends amid AI diffusion using National Pension subscriber data and the Economically Active Population Survey. First, the decline in youth employment remains concentrated in industries with high AI exposure. Between June 2022 and June 2026, youth employment fell by 285,000, with high-exposure industries accounting for 268,000, or 94%, of the decline. Sharp declines occurred in information services (-31.4%), publishing (-27.4%), computer programming (-16.6%) and professional services (-11.6%). Employment among workers in their 50s continued to rise in these same industries. This pattern is consistent with the persistence of the “seniority-biased technological change” documented by Han and Oh (2025).

Second, youth employment has declined less in industries where AI is used to augment human work. Industries with a higher share of AI use for automation saw larger declines in youth employment. No comparable pattern emerged for augmentation, where AI assists rather than replaces human work. These findings suggest that AI’s employment effects may depend on how it is used, not simply on whether it is adopted.

Third, unemployment has risen more among young university graduates, who tend to be more exposed to AI. Since November 2022, the unemployment rate has averaged 7.0% among young people with a bachelor’s degree or higher, compared with 5.4% among those without one—a gap of 1.6 percentage points. Before then, the two groups had similar unemployment rates.

Fourth, the decline in youth employment reflects not only reduced hiring but also increased exits from employment. More young people leaving high-exposure industries have also entered unemployment. The challenge therefore extends beyond finding a first job: young people may also face greater difficulty staying employed and accumulating experience and skills.

These patterns do not establish that AI caused the recent weakness in youth employment. The reversal of pandemic-era over-hiring, a growing preference for experienced hires, weaker in-house training and the spread of remote work may all have contributed. Rather than acting as a standalone shock, AI may be accelerating pre-existing changes in hiring and work practices.

The recent decline need not imply a persistently weaker employment outlook for young people. Their active use of AI and ability to adapt to new technologies could make them major beneficiaries of productivity gains and new jobs over time. Policy should focus on rebuilding career ladders rather than preserving existing entry-level jobs, enabling young people to gain experience and skills while working with AI. Priorities include work-based apprenticeships, support for firms’ investment in training and mentoring, and a review of whether policy incentives strike an appropriate balance between investment in technology and human capital.