Abstract

This paper investigates the labor market effects of artificial intelligence (AI) on the German labor market, thereby extending the approach of Brynjolfsson et al. (2025) to the German context. Using individual-level administrative data, we examine how employment patterns differ over time and across occupations with varying exposure to AI technologies. Additionally, we test how much the observed pattern depends on individuals' characteristics, especially workers' age and tenure. Furthermore, we look into heterogeneity across regional and institutional contexts, such as union density and local labor market tightness, reflecting outside options on the labor market.