AI Adoption in Subnational Governments: Evidence-Based Findings from Brazil

• In Event: AI in Public Sector Service Delivery

Abstract

This article challenges the assumption that fiscal resources drive AI institutionalization in subnational governments, demonstrating instead that organizational capability, specifically the endogenous capacity to develop and adapt digital solutions, has a strong association with advanced AI use in public organizations. Using a mixed-methods design, the study combines original primary data collected through the Freedom of Information Act from all 27 states and 337 municipalities with more than 100,000 inhabitants. Grounded in the Resource-Based View theory (RBV), the analysis tests hypotheses related to fiscal, administrative and technological capacity, and intergovernmental coordination. The findings indicate that fiscal capacity is associated with the initial adoption of AI but do not explain its institutionalization. Similarly, bureaucratic stability shows no significant relationship with the diffusion of technology. In contrast, endogenous technological capabilities emerge as key factors of AI maturity. The study contributes to the literature on digital governance by showing that, in the Brazilian federal context, the institutionalization of AI depends less on fiscal resources than on the organizational capacity to develop and adapt technological solutions in-house. The findings point to a fragmented pattern of innovation, raising concerns about growing digital disparities and the need for coordinated intergovernmental policy responses. Two broader arguments follow: first, that RBV-informed research on public sector digitalization should treat resource endowment and organizational capability as empirically distinct constructs; and second, that vertical isomorphism cannot be assumed in federal systems marked by structural inequality, a finding with direct implications for how equity in digital governance is theorized and measured.

Author

• Otavio Ritter, Fundação Getulio Vargas
  Presenting Author