AI Workforce Transformation Wiki

Not yet human-reviewed

This source record was written automatically by the research pipeline and has not yet been reviewed by the maintainer or integrated into a topic page. Treat it as a provisional summary and verify claims against the original source.

Robots & AI exposure and wage inequality: a within occupation approach

Citation. Jaccoud, F. (2025). Robots & AI exposure and wage inequality: a within occupation approach. Eurasian Business Review, 15, 1035–1090. https://doi.org/10.1007/s40821-025-00306-w

Summary. This study relates occupational exposure to AI and industrial robots to within-occupation wage inequality across 19 European countries. Fixed-effects estimates associate greater AI exposure with wider wage dispersion, especially in the upper half of the distribution, while greater robot exposure is associated with lower dispersion, particularly in the lower half. The ingested PDF is the UNU-MERIT working-paper version; its published journal record provides the canonical provenance.

Key claims

  • Occupations with higher AI exposure have greater within-occupation wage inequality, with the association concentrated in p90/p50 wage gaps.
  • Occupations with greater robot exposure have lower within-occupation wage inequality, particularly in p50/p10 wage gaps.
  • The contrasting associations are consistent with AI exposure being linked to high-skill cognitive task content and robot exposure to routine physical task content.

Evidence & limitations

  • The analysis uses EU Structure of Earnings Survey data for 19 European countries, Webb (2020) occupation-level AI and robot exposure scores mapped to ISCO-08, and fixed-effects models of 2-digit occupational wage gaps.
  • Exposure measures describe task overlap rather than actual adoption, and the study explicitly does not make causal claims; the survey excludes self-employment, family workers, firms with fewer than 10 employees, and public administration.
  • The ingested 62-page Drive PDF is UNU-MERIT Working Paper 2025-013 (DOI 10.53330/EAJL3597); Crossref and the publisher identify a 2025 peer-reviewed version of record with the canonical DOI above, so binary equivalence between the Drive working-paper file and the published article was not claimed.

Feeds