Epidemic Effects in the Diffusion of Emerging Digital Technologies: Evidence from Artificial Intelligence Adoption

The properties of emerging, digital, and general-purpose technologies make it hard to empirically capture adoption by firms and theoretically its determinants. This study builds on epidemic models of inter-firm technology diffusion and adds concepts of relational embeddedness from social network theory. We apply our model of AI adoption to a broad web-based data sample of more than 380,000 firms in Germany, Austria, and Switzerland. To this end, we train a transformer-based language model to identify firm-level AI adoption from over 1.1 million websites and replicate our econometric analysis for representative firm-level survey data from Switzerland. We show that AI adoption can be contagious under three circumstances: 1) When companies in industrial and regional hot-spots associated with the production of AI knowledge are subjected to emulation pressure. 2) When companies foster a high intensity of strong direct ties transmitting deep AI knowledge. 3) When companies are heavily embedded in the AI knowledge network. Overall, our findings indicate a highly clustered pattern of diffusion and a rather closed system of AI adopters that is likely to hinder broader diffusion. Finally, we discuss how innovation policy may support diffusion activities to counterbalance the narrow localization of AI usage provoked by cluster-based policies.

Dahlke, Johannes, Mathias Beck, Jan Kinne, David Lenz, Robert Dehghan, Martin Wörter und Bernd Ebersberger (2024), Epidemic Effects in the Diffusion of Emerging Digital Technologies: Evidence from Artificial Intelligence Adoption, Research Policy 53(2)