Brazil sets age-based rules for AI use in schools and universities

Brazil has introduced national guidelines governing the use of artificial intelligence (AI) across schools, universities and other educational institutions, setting age-based restrictions for students and requiring human oversight of academic decisions involving the technology.

The National Education Council (CNE) adopted Resolution CNE/CP No. 1 on Nov 11, establishing rules for the use of AI across all levels and stages of Brazilian education. The guidelines apply to public and private educational institutions and cover teaching, learning, research, assessment and institutional administration.

The resolution states that AI must be used for legitimate educational purposes, with appropriate human supervision, protection of personal data and safeguards for students' rights. It also makes clear that automated systems cannot replace teachers or assume institutional responsibility for educational decisions.

Restrictions on AI use by younger students

The guidelines introduce different conditions for AI use according to students' age and stage of education.

In early childhood education, children are prohibited from using AI systems, whether independently or with a teacher's assistance. Educators may use the technology for planning and educational management, while accessibility tools may be used for the specific purpose of removing learning barriers.

For the initial years of primary education, AI-based resources may be used only through activities planned and conducted by educators. Students cannot independently access AI systems, and interaction with generative or conversational AI tools is prohibited even under teacher supervision.

In the later years of primary education, generative AI tools may be used only in activities mediated by educators. Direct, independent or unsupervised student use is prohibited.

The resolution also calls for AI literacy to be integrated into basic education, enabling students to understand how these systems work, assess their limitations and identify potential biases and misinformation.

Teachers retain control over assessments

The guidelines place limits on the role of AI in evaluating students' work. Automated systems may assist with the preliminary correction of essays and other written assignments, but they cannot independently determine marks or academic outcomes.

Teachers must read and assess students' work themselves, while any AI-generated suggestions must remain subject to meaningful human review. The resolution also prohibits decisions affecting students' academic progression, certification or disciplinary status from being made exclusively through automated systems.

It further states that AI-detection tools cannot be the sole basis for imposing academic or disciplinary penalties on students.

The rules also prohibit practices such as emotional surveillance of students, continuous biometric monitoring and the commercial exploitation of educational data. Student work cannot be used to train third-party AI models without specific authorisation, while content produced by children and adolescents is excluded from this exception.

Universities must establish their own AI policies

For higher education institutions, the resolution allows AI to be incorporated into teaching, research and academic administration, subject to institutional planning and human oversight.

Universities are required to establish their own policies covering AI use in teaching, research, extension activities and academic management. These policies are expected to address academic integrity, transparency, intellectual responsibility and data protection.

The guidelines also address the use of AI in research. Researchers must disclose the technology's contribution when it has played a relevant role in activities such as writing, data analysis, literature reviews or programming. The use of AI does not transfer authorship or responsibility for research findings to the technology.

Institutions get 12 months to comply

Educational institutions and education systems have 12 months from the resolution's publication to bring their policies, practices, governance arrangements and existing contracts in line with the new requirements.

However, the prohibitions specified in the resolution's Article 41 apply immediately. These include restrictions on emotional surveillance, certain forms of biometric monitoring, exclusively automated decisions with significant academic consequences and the use of children's educational data to train third-party AI models.

The National Education Council will monitor implementation in coordination with education systems, institutions, researchers and other stakeholders.

The resolution establishes a national framework for AI use in education while leaving institutions responsible for developing policies suited to their academic and administrative requirements.