Toward AI-Ready Auditing

A Model Context Protocol (MCP) Network for Oversight Institutions

Autores

  • Adriano Marabuco de Albuquerque Lima TCE/PE

Palavras-chave:

Model Context Protocol (MCP), Artificial Intelligence Agents, Public Sector Auditing, Digital Government, Accountability, Interoperability

Resumo

O trabalho propõe uma rede federada de servidores Model Context Protocol (MCP) para instituições de controle, permitindo a exposição segura e padronizada de dados e fluxos de auditoria. A arquitetura facilita a interação de agentes de IA sob rigorosos controles de autenticação e governança, promovendo outputs auditáveis e reprodutíveis. São detalhados um modelo de referência, fluxos explicáveis e salvaguardas de governança, além de um caso ilustrativo em auditoria de folha de pagamento. A proposta visa superar iniciativas isoladas, promovendo interoperabilidade e transparência. O enfoque é na construção de um ecossistema de auditoria digital centrado no cidadão e preparado para IA.

Biografia do Autor

Adriano Marabuco de Albuquerque Lima, TCE/PE

Ph.D. (in progress) and an M.Sc. in Computer Science from the Federal University of Pernambuco (UFPE). He also holds degrees in Law (USP) and Business Administration (FGV/EAESP), with specializations in Public Law and Machine Learning. His research focuses on artificial intelligence, time series forecasting, and auditing applications, with publications in Procedia Computer Science and leading Brazilian computing conferences (SBSI, ENIAC). Professionally, he serves as an External Control Analyst at the Court of Accounts of Pernambuco (TCE/PE) and has prior experience as an Internal Control Auditor in São Luís. His interdisciplinary background combines computing, management, and law, with a particular interest in applying AI and data science to enhance public sector accountability.


Lattes: http://lattes.cnpq.br/5477365718517046

E-mail: adrianomarabuco@tcepe.tc.br

Referências

DAWES, S. Interagency information sharing: Expected benefits, manageable risks. Journal of Policy Analysis and Management, v. 19, n. 2, p. 253–277, 2010.

JANSSEN, M.; KUK, G. Big and open linked data (BOLD) in government: A challenge to transparency and accountability. Government Information Quarterly, v. 33, n. 4, p. 563–567, 2016.

KARPAS, E. et al. What can AI agents do? A survey. Proceedings of AAAI, 2022.

OECD. Digital Government Review. Paris: OECD Publishing, 2020.

POWER, M. The Audit Society: Rituals of Verification. Oxford: Oxford University Press, 2021.

ZHANG, J.; YANG, L. Detecting anomalies in public sector data with machine learning. ACM Transactions on Management Information Systems, v. 12, n. 3, p. 1–18, 2021.

Downloads

Publicado

2026-07-17

Edição

Seminário Internacional O Futuro da Auditoria Pública

Seção

Eixo 4 - Inteligência Artificial e transformação digital na auditoria pública