Inteligência artificial na pesquisa científica brasileira: uma análise comparativa de documentos oficiais

This study comparatively analyzes the documents Artificial Intelligence in Research and Funding: Challenges and Opportunities (CAPES) and CNPq Portaria [administrative ruling] No. 2,664, of March 6, 2026, with the aim of examining how Generative Artificial Intelligence is regulated in Brazilian scientific research. The corpus consists of these two official documents, subjected to qualitative, documentary, descriptive, and comparative research. The analysis was organized around six axes: normative nature and scope of application; authorship, responsibility, and human oversight; transparency and disclosure obligations; scientific integrity, infractions, and sanctions; data protection, confidentiality, and scientific secrecy; and training, competencies, and institutional culture. The results indicate convergence in the affirmation of human responsibility and divergence regarding the degree of mandatory enforcement, level of detail, and sanctioning force. Silences were also identified concerning the local operationalization of guidelines and inequalities in access to Generative Artificial Intelligence tools. It is concluded that the governance of Generative Artificial Intelligence in Brazilian science is in a process of consolidation, combining guidance, regulation, and gaps that have yet to be overcome.