The geography of artificial intelligence in European regions: innovation, exposure and use

This paper maps the geography of artificial intelligence (AI) innovation, exposure and use across European Union (EU) NUTS-2 regions using six complementary indicators. Findings reveal a spatial paradox: while AI exposure and worker-level use are relatively widespread, innovation and firm-level adoption remain concentrated in high gross domestic product (GDP), human capital-intensive hubs. A persistent implementation gap separates structural exposure from productive use, driven by institutional, organisational and capability-related frictions. Income and human capital are strongly associated with AI activity, while trade openness shows a negative relationship. Results point to the need for place-sensitive policies combining support for leading regions with targeted capability-building in lagging ones.