The role of human-capital in artificial intelligence adoption

We study the role of human-capital on the early adoption and diffusion of Artificial Intelligence (AI) technologies. We do this using recently released EUROSTAT data on AI-adoption across European countries and industries in 2021. We find that pre-existing differences in human-capital are a key driver of differences in AI-adoption. Our estimates suggest that one-third-to half of AI-adoption is linked to differences in available human-capital across country-industry cells. This relationship holds even when we account for country and industry specific factors. Moreover, higher clustering of university educated workers in firms appears crucial for AI-adoption. This effect is particularly strong for the following AI technologies: (i) those assisting the automation of processes and (ii) those that enable physical movements of machinery. Within firms, human-capital fosters the use of AI in processes associated with production, management and logistics. Despite documenting the automative nature of the AI technologies adopted, we find no evidence for them leading to employment declines (at least by 2021).