Artificial Intelligence, Algorithmic Transparency and Financial Citizenship in Brazil: empirical mapping of adoption, internal governance and consumer vulnerabilities

Abstract

This article examines the intersection of artificial intelligence adoption in Brazil’s National Financial System, emerging algorithmic transparency regimes, and the impact on the constitutive dimensions of financial citizenship. The research combines six phases: institutional mapping by cross-referenced evidence; functional analysis of fifty financial applications; public complaint mining using natural language processing; documentary examination of current regulation; comparative analysis of international governance; and operationalization of twelve proxy indicators of financial citizenship. Results indicate that AI adoption is substantially more advanced than the regulatory framework acknowledges, with widespread use of machine learning in credit and customer service and growing experimentation with large language models. Internal governance is heterogeneous, concentrated in large banks and largely absent in fintechs. Dependence on foreign technology providers is massive and unmapped, compromising the State’s auditing
sovereignty. The hypotheses examined reveal structural misalignment between regulatory perception and consumer
experience, evidencing a normative vacuum deepened by the absence of a general artificial intelligence statute in Brazil and by specific gaps in Law No. 15.252/2025. The article proposes a layered flexible regulatory architecture, a monitoring indicators table, and a research agenda for open frontiers.