Between guiding and prescribing: the institutionalization of generative AI in Brazilian public universities

Abstract

Institutional documents on generative artificial intelligence (GenAI) have proliferated, mobilizing apparently consensual formulas such as “responsible use” and “best practices”, yet how these expressions are translated into concrete rules in Brazilian Higher Education remains an open question. This study compares normative texts from six public universities: the Federal University of Minas Gerais (UFMG), São Paulo State University “Júlio de Mesquita Filho” (UNESP), the State University of Campinas (UNICAMP), the Federal University of Bahia (UFBA), the Federal University of Goiás (UFG), and the Federal University of Ceará (UFC), approaching them not as mere repositories of information, but as governance instruments that organize conduct and allocate responsibilities within the historical context of the growing datafication of intellectual work. The qualitative documentary analysis identified four components in each text — general principles, practical requirements, control mechanisms, and formative guidance — and also assigned an index of normative explicitness. The findings point to divergences on three fronts: the extent to which each document formalizes obligations, the extent to which it makes prohibitions explicit, and who decides on the use — the central administration or the faculty member in the classroom. The study concludes that, beneath a shared lexicon, universities sustain distinct regulatory architectures, outlining a discreet dispute over the regime of academic accountability in relation to GenAI.

Keywords: Generative artificial intelligence. Academic governance. Public universities.