How does Artificial Intelligence (AI) affect the task content of jobs, and how do workers adjust to AI's diffusion? To answer these important questions, we combine novel patent-based measures of exposure to pre-generative AI and robots with individual survey data on job tasks and administrative data on worker careers. Consistent with prior studies, we find that robots have reduced routine tasks. In sharp contrast, early AI technologies have reduced abstract tasks like information gathering for all workers. The resulting task reallocation is much more pronounced for high-skilled workers: they switch to other abstract tasks like teaching and training and to complex routine tasks like monitoring processes within the same occupation. While displacement effects are small, AI-exposed workers are more likely to switch employers, often to less exposed industries. We document modest wage declines for incumbent low-skilled workers as well as lower entry wages and fewer employment opportunities in AI-exposed industries.

We study how export shocks affect earnings inequality between workers. While existing approaches document heterogeneous effects, they provide limited insight into the full distribution of effects across workers and into the consequences for earnings inequality. We use a generalized random forest to estimate individualized earnings responses to exports. We find substantial dispersion in gains from exports, with skill and firm characteristics explaining a significant share, but leaving considerable variation accounted for by demographic and career-related factors. Despite this heterogeneity, the aggregate impact on inequality is modest, with earnings gains peaking in the upper middle of the earnings distribution. Our results show that accounting for high-dimensional heterogeneity is crucial for understanding the distributional effects of trade.