Regulatory Triggers, Technical Answers: Mapping Laws to Explainable AI Controls in the Brazilian Financial Sector

• José Eduardo Vaz ULBRA
• Pedro Henrique Silva de Oliveira UFPel
• Felipe Bertoglio UFRGS
• Luiz Nystrom Unicesumar
• Cleber Kazanowski PUCRS

DOI: https://doi.org/10.5753/eramiars.2025.16786

Resumo

This paper analyzes the regulatory framework for Explainable AI (XAI) in Brazil's financial sector, focusing on the period before the Bill Draft 2.338/2023. We argue that while no unified law on explainability exists, a functional governance structure has emerged through sector-specific regulations from the National Monetary Council (CMN) and the Central Bank of Brazil (BCB). By mapping relevant resolutions, we identify a focus on auditability and risk management in statistical models, rather than a broad right to explanation for consumers. We conclude that Brazil’s evolving AI regulation must reconcile these sectoral rules with a broader, principle-based approach to ensure transparency and protect fundamental rights.

Referências

Brasil (1942). Decreto-Lei 4657/1942. Lei de Introdução às normas do Direito Brasileiro.

Brasil (2019). 23.610/2019. RESOLUÇÃO No 23.610, DE 18 DE DEZEMBRO DE 2019.

Long, P. and Solaimani, S. (2025). Beyond the black box: operationalising explicability in Artificial Intelligence for financial institutions. International Journal of Business Information Systems, v. 49, n. 1.

Macpherson, C. B. (1990). The political theory of possessive individualism: Hobbes to Locke. 13. impr ed. Oxford: Oxford Univ. Press.

Ortiz, G. A. (2005). Sucessos e fracassos da regulação. Revista Eletrônica de Direito Administrativo Econômico, n. 3, p. 1–13.

Scholze, S. H. C. and Wimmer, M. (2009). A Regulação Das Telecomunicações No Brasil: Passado, Presente E Desafios Futuros. Desafios da regulação no Brasil. 1. ed. Brasília, DF: PRO-REG. p. 155–182.