Brazil establishes new AI guidelines for education

The National Council of Education (CNE) has established new guidelines for the use of artificial intelligence in Brazilian education. The regulations categorize AI applications into four risk levels: low, moderate, high, and excessive.

High-risk applications, which include automated grading of academic assessments, biometric proctoring, and selection or certification systems, are prohibited from operating without significant and continuous human supervision. These high-risk uses require prior algorithmic impact assessments and data protection reports. Excessive-risk applications, such as social scoring, emotional surveillance, and exclusively automated decisions regarding student promotion or retention, are strictly forbidden.

In contrast, low-risk uses like organizing didactic materials or supporting accessibility are permitted with minimal transparency requirements. Moderate-risk tools, such as tutoring systems and formative feedback, require mandatory human review and institutional records. The guidelines emphasize that while AI can support learning, accessibility, and management, it must remain subordinate to educational purposes and cannot replace human judgment or responsibility.