Artificial Intelligence And Inequality : Evidence from OECD Countries

Sammanfattning

This paper explores the relationship between the rise of technology related to artificial intelligence and income inequality in OECD countries. To measure income inequality, the study uses data on the income shares of different population percentiles and employs AI patents as a measure of innovation. Additionally, it incorporates Information and Communication Technology (ICT) patents to assess how ICT compares to AI in terms of its relationship with income inequality. Studying long-run changes 2005-2020 in a long-difference regression, this study finds a positive, albeit insignificant, relationship between AI intensity and inequality. Furthermore, the study examines the relationship between AI intensity and employment share by skill, revealing a positive relationship for high-skilled individuals and a negative one for low-skilled individuals. Medium-skilled workers are found to be relatively worse off in countries with high AI intensity, especially in the context of low trade union presence. The results this paper got align with the Skill-biased technological change in terms of the AI innovation whereas it aligns with the job polarization and u-shaped or task-based framework in case of the ICT innovation. This suggests the importance of studying the labor market and how it gets influenced by the innovation.