The project examines how AI policy shocks translate latent occupational exposure into observable employment restructuring. It begins from the premise that technological diffusion does not necessarily produce labor market adjustment: exposure becomes consequential only when market forces and institutional conditions activate the overlap between what technology can do and what workers actually do. Rather than treating occupations as fixed labels, the project focuses on the task-level alignment between technological capabilities and job activities. Combining large language model assessments with detailed occupational task profiles, it identifies who builds AI, who works with it, and who is more likely to be displaced or reconfigured. Using firm-level vacancy data from China and staggered city-level AI policy interventions, the project traces how policy activates these exposure structures, distinguishes destructive from transformative adjustment, and evaluates heterogeneous effects across policy instruments and their combinations. It further explains these patterns through two linked mechanisms—firms’ technology adoption decisions and regional technological endowments—and, in doing so, advances a policy framework for how AI-related policy tools can be more effectively configured to support employment transformation and more precise labor market governance.

Researchers

• Yingzi Qu (School of International Studies | University of Trento)

           Supervisor: Mauro Caselli (School of International Studies | University of Trento)

Funding

HORIZON-MSCA-2024-PF-01