Artificial Intelligence and Employment in the OECD: A Panel Data Analysis

Abstract

This thesis examines the relationship between artificial intelligence (AI) and employment across 35 OECD countries from 2013 to 2022 using panel data analysis. By employing ordinary least squares (OLS) and fixed effects regression models, the study explores how AI, proxied through AI patent activity and private investment, affects total employment, with a focus on demographic (gender and age) and sectoral heterogeneity. The results illustrate a positive association between AI and employment, consistent with the labor augmentation theory. Employment growth was observed across all demographic groups and economic sectors. However, the employment gains were mostly concentrated among the middle-aged workers and in the service sector. While evidence indicates that AI can act as a complement to human labor and create new jobs, the analysis is constrained by data limitations and potential endogeneity concerns. The findings emphasize the importance of inclusive, forward-looking policies that promote skill development and equitable AI adoption across sectors and groups.