Center for Artificial Intelligence and Public Governance (CIA-GP)

• Grant number: 25/07136-2
• Support Opportunities: Research Grants - Science Centers for Development
• Start date: February 01, 2026
• End date: January 31, 2031
• Field of knowledge: Applied Social Sciences - Law
• Principal Investigator: Marcelo Fantinato
• Grantee: Marcelo Fantinato
• Host Institution: CENTRO ESTUDOS INTELIGENCIA ARTIFICIAL APRENDIZADO MAQUINA/USP
• Principal investigators: Alessandro Hirata ; Cristina Godoy Bernardo de Oliveira ; João Paulo Cândia Veiga ; José Carlos Vaz ; Juliano Souza Albuquerque Maranhão ; OMAR CASSIM NETO ; Sarajane Marques Peres
• Associated researchers: Cíntia Rosa Pereira de Lima ; Daniel de Angelis Cordeiro ; Edelcio Gonçalves de Souza ; Jorge Alberto Silva Machado Villanueva ; José de Jesús Pérez Alcázar ; Karina Valdivia Delgado ; Leandro Augusto da Silva ; Marcelo de Souza Lauretto ; Norton Trevisan Roman ; Rafael Meira Silva ; Renata Mendes de Araujo ; Renata Wassermann ; Silvia Inês Dallavalle de Pádua ; Theo Lovizio de Araujo

Abstract

The Center for Artificial Intelligence and Public Governance (CIA-GP), proposed by the University of São Paulo in partnership with the São Paulo State Government's Civil House Secretariat, aims to address the growing complexity of normative governance through the development and application of advanced techniques in Artificial Intelligence (AI), process mining, and computational law. This initiative seeks to significantly enhance the analysis, processing, and monitoring of regulatory proposals within the public sector. The State of São Paulo faces a high volume of normative acts that demand precision, standardization, and traceability-essential elements for ensuring efficient, transparent, and legally sound governance. Despite the existing digitization infrastructure provided by the Electronic Information System (SEI), a critical gap remains in the integration of advanced analytical tools. This gap results in limited predictability, high operational costs, and increased risks of legal inconsistencies. CIA-GP proposes an innovative and interdisciplinary approach, combining expertise in computer science, information systems, law, public administration, and political science. Its main research lines include the computational structuring and formalization of legal norms, the development of domain-specific language models (legal LLMs), the application of advanced process mining techniques, and the implementation of robust algorithmic governance systems with a strong emphasis on explainable AI (XAI). These methods will enable the automatic detection of normative inconsistencies, the anticipation of regulatory impacts, and more reliable and transparent decision-making processes. Among the specific objectives are the development of automated systems for regulatory analysis, platforms for visualization and optimization of administrative workflows, predictive models for regulatory impact, and tools for the automated summarization of technical reports. The strategic partnership with the Civil House Secretariat ensures privileged access to real-world data and governmental environments, which are essential for the immediate and practical validation of proposed solutions. The implementation of CIA-GP is expected to yield significant short- and long-term outcomes, including major improvements in operational efficiency, technical quality, and digital public governance. Furthermore, the Center will contribute to establishing the State of São Paulo as a national and international reference in data-driven, evidence-based regulatory innovation, thereby expanding its social, economic, and institutional impact. (AU)