Impact of AI-driven reorganization on scientific expertise and skill acquisition

Determine how the reallocation of research resources toward human capital, larger and more diverse teams, and expanded task scope associated with the adoption of modern artificial intelligence methods in scientific projects affects the expertise required of scientists and the long-run evolution of scientific skill acquisition.

Background

The study analyzes a comprehensive corpus of research proposals to one of the largest international biomedical and health funding agencies, identifying when and how modern AI is incorporated into planned research. It finds modest short-run scientific returns to AI adoption but substantial effects on research organization: AI-enabled projects shift budgets toward human capital, involve larger teams, and span a broader set of tasks.

These organizational changes are consistent with AI functioning as a general-purpose technology that first induces reallocation and reorganization before productivity gains materialize. Given these shifts, the authors highlight uncertainty about how the required expertise of scientists and the pathways of skill acquisition will evolve as AI capabilities mature and integration deepens.

References

How these changes affect the expertise required of scientists and the long-run evolution of skill acquisition remains an open question (Acemoglu et al., 2026; Garicano & Rayo, 2025).

— Artificial Intelligence in Science: Returns, Reallocation, and Reorganization  (2603.27956 - Hosseinioun et al., 30 Mar 2026) in Conclusion