An old familiar song? Assessing the artificial intelligence divide among the regions of the European Union and its connections with digital divides

This paper explores the use of artificial intelligence (AI) tools by firms across the regions of the European Union. Using the multivariate technique of factor analysis on a set of AI variables, we assess regional levels of AI use and identify both leading and lagging regions. Within countries territorial disparities are also measured. In addition, we explore the extent to which AI use is linked to regional levels of digitalization. Using spatial econometric models, our findings suggest that regional levels of AI use by firms are explained by past usage of digital tools in a region, i.e., by factors related to the second level of the digital divide. No significant relationship is found either for access or outcomes. Furthermore, what neighboring regions do also play a role. Results indicate the existence of complementary effects in AI use between a region and neighbors; meanwhile the relationship would be negative between a region's use of AI and neighbors' digital use. • Danish and Belgian regions are at the forefront of AI use by firms. • Meanwhile, Romanian and Bulgarian regions lag behind. • Regional levels of AI use can be explained by firms' past usage of digital tools and their neighbors' behavior. • Regions whose neighbors exhibit high levels of AI use tend to show similarly high levels of usage. • However, neighbors’ digital use is negatively associated with AI use, suggesting competition effects between territories