How should macroeconomists think about AI and the labor market?

How does generative artificial intelligence (AI) affect labor markets? Existing models of skill-biased and routine-biased technical change do not fully explain recent worker-level evidence: early employment effects are concentrated among highly educated, entry-level workers in prediction-intensive occupations, appear chiefly in hiring rather than wages, and remain limited in the aggregate. This paper develops a conceptual framework that integrates task-based automation with a decision-theoretic account of AI as cognitive augmentation. AI reshapes the allocation of tasks by changing how firms combine prediction, judgment, and expertise. The framework connects these organizational adjustments to aggregate labor-market outcomes and sets out a research agenda on AI’s effects on employment, wage inequality, and economic growth.