In the rapidly evolving landscape of higher education, artificial intelligence (AI) tools like ChatGPT and similar generative models have become ubiquitous among students and faculty. Yet, with great power comes the need for guidelines to ensure academic integrity. A recent survey by researchers at the University of São Paulo (USP) reveals that only 12 Brazilian universities have published formal rules or manuals governing AI use in academic work, marking a pivotal step toward ethical integration of this technology. This development comes amid widespread adoption, with surveys indicating that up to 70% of Brazilian students incorporate AI into their study routines.

These institutions, primarily public federal and state universities, are responding to concerns over plagiarism, skill erosion, and fairness by emphasizing transparency, human oversight, and pedagogical value. As Brazil's higher education sector—home to over 2,500 institutions—grapples with AI's dual role as both enabler and potential disruptor, these pioneering guidelines offer a blueprint for others.

🌐 The USP Survey: Spotlight on Brazil's AI Governance Pioneers

The landmark study from USP's Institute of Advanced Studies (IEA-USP), led by the "Governança de Agentes de IA" group, scanned Brazil's vast higher education network. Out of thousands of universities, just 12 have concrete, published directives. While the exact list isn't exhaustive in public reports, prominent examples include Universidade Estadual Paulista (Unesp), Universidade Federal de São Paulo (Unifesp), Universidade Federal Fluminense (UFF), Universidade Federal da Bahia (UFBA), Universidade Federal do Ceará (UFC), Universidade Federal de Mato Grosso do Sul (UFMS), Universidade Federal do Rio de Janeiro (UFRJ), Universidade Federal da Paraíba (UFPB), Universidade Federal de Alagoas (Ufal), Universidade Federal de Uberlândia (UFU), and Universidade Federal de Goiás (UFG).

This scarcity underscores a broader trend: nearly half of Brazil's 69 federal universities (43%, or 30 institutions) either have guidelines or are debating them, per a O Globo analysis. The slow pace reflects bureaucratic hurdles and the technology's swift evolution, but experts like Jonas Gonçalves from the USP group advocate for pedagogical governance over outright bans, prioritizing training in AI's ethics and limits.

Spotlight: São Paulo's Power Trio – USP, Unicamp, and Unesp

The state universities of São Paulo are at the forefront. Unesp launched a comprehensive guide for undergraduates, categorizing AI uses into permitted, prohibited, and conditional. Permitted activities include text translation, paraphrasing, summarization, grammar checks, and generating educational visuals like images or mind maps. Prohibited: submitting AI-generated work as original without declaration or using it in unauthorized exams.

USP and Universidade Estadual de Campinas (Unicamp) are finalizing similar protocols, mandating detailed disclosures: tool name, version, prompts used, and application phases. Professors must collaborate with students on usage, review AI outputs, and declare their own applications in teaching materials. This collaborative approach aims to foster critical thinking while harnessing AI's potential.

Unicamp's "Diretrizes para um uso ético e responsável da IA Generativa" provides practical advice for researchers, downloadable as a PDF for reference.Read the full Unicamp guidelines here.

Federal Universities Lead with Diverse Approaches

Federal institutions showcase varied strategies. Unifesp's December 2025 post-graduate rules require explicit declarations in final submissions, with peer review to validate AI-assisted content. UFF's guide permits AI for topic structuring and bibliography suggestions but demands prompt inclusion to verify authenticity.

• UFBA: Professors define per-discipline rules; students must detail tools, prompts, and outputs to demonstrate learning.
• UFC: Bans AI for core research sections like methods and results; mandates Turnitin checks.
• UFMS: Allows data analysis and simulations with student evaluation of outputs.
• UFRJ: Treats unauthorized use as plagiarism-equivalent misconduct.

UFBA's ethical guide is a model resource.Access UFBA's full guide.

Core Principles: Transparency and Human Supremacy

Across all guidelines, two tenets dominate: transparency and human oversight. Students must cite AI like any source—specifying model (e.g., GPT-4o), date, and prompts—for reproducibility and accountability. Professors redesign assessments (e.g., oral defenses, projects) to prioritize synthesis over generation.

AI shines in supportive roles: literature reviews, idea brainstorming, or visualizations. But generating original analyses or fabricating data is taboo, as it undermines research integrity. Detection tools like Turnitin are recommended, though cautioned against false positives.

Challenges and Risks in Brazil's AI Academic Landscape

Brazilian educators worry about cognitive atrophy, where over-reliance on AI hampers critical thinking. A 2025 survey found 85% of students across 166 universities have used generative AI, with 25% for full task completion. Plagiarism risks escalate with undetectable outputs, prompting calls for updated policies.

Cultural context matters: In a nation with stark educational inequalities, AI could widen gaps if access to premium tools varies. Faculty training lags, with many needing upskilling on ethical deployment.

Student and Faculty Perspectives: Voices from the Ground

Students appreciate AI for efficiency—70% use it daily for summaries or translations—but fear unfair penalties from imperfect detectors. Professors like those at USP emphasize dialogue: "IA deve ser combinada," integrating it as a tool, not crutch.

Stakeholders advocate balanced views: AI augments, doesn't replace, human intellect. Initiatives like Unesp's workshops train users on ethical prompts and bias awareness.

Global Context: Brazil in Step with Worldwide Trends

Brazil aligns with global shifts. Harvard and Stanford mandate disclosures; the EU's AI Act influences higher ed ethics. Yet, Brazil's federal focus contrasts private sector lags. MEC's 2026 referencial promotes equity, echoing UNESCO's guidelines.

Photo by Gustavo Sánchez on Unsplash

Future Outlook: Toward National Standards?

With 12 trailblazers, momentum builds. Experts urge MEC-led national frameworks, faculty development, and AI literacy curricula. By 2030, expect widespread adoption, preparing graduates for AI-driven jobs. For now, these rules safeguard integrity while embracing innovation.

Explore opportunities in Brazil's higher ed sector via AcademicJobs Brazil listings.