This paper aims to assess the determinants of artificial intelligence (AI) adoption in Brazilian enterprises, evidence that is crucial for understanding and accelerating its diffusion to boost innovation and productivity.
The authors use microdata from Brazilian firms. The empirical strategy consists of standard, ordered and multivariate probit models to identify the patterns of adoption, intracompany diffusion and intensity of use.
Results suggest that company size positively influences AI adoption and use due to greater resources and risk tolerance. Environmental variables also play a significant role, with digitized environments increasing adoption likelihood due to network effects, while competition intensity seems to have a nonlinear effect, initially stimulating adoption but decreasing it at higher competitive intensity. Existing digitization infrastructure of adopting firms promotes AI adoption and use.
Public policies should stimulate workers’ skills, as well as encourage the digitization even of less sophisticated but complementary technologies for AI. This is reinforced by the existence of network effects. In addition, competition authorities have a very important role in promoting healthy competition because it represents an indirect stimulus for AI adoption.
This contribution is one of the few empirical analyses of the drivers of AI at the firm level based on an extensive microdata sample, and to the best of the authors’ knowledge, the first that contemplates intrafirm diffusion, usage intensity and the role of network effects.
