Abstract This study investigates how French firms use Artificial intelligence (AI), drawing on a uniquely detailed and nationally representative dataset that reports both the specific AI technologies implemented and the business functions in which they are deployed in 2020 and 2022. Evidence on sectoral use rates, the interdependencies between AI technologies, and their applicability across business functions shows that three technologies lie at the core of the AI paradigm and exhibit a more general-purpose nature: Machine Learning Data Analysis, Text Mining, and Automation & Decision Support. This pattern supports the view that AI is not a monolithic technology but a system of heterogeneous technologies with varying degrees of generality. Regression analyses further demonstrate that firms using different AI technologies are far from homogeneous. These results have two important implications. First, diffusion strategies should recognize the distinct characteristics of individual AI technologies. Second, AI use should be treated as part of a broader strategy involving multiple technologies that are interdependent and have different degrees of applicability across business functions.
Calvino et al. (Mon,) studied this question.