Brazil Moves Toward First AI-in-Education Rules, With Ban on Automated Essay Grading

Overview

• An initial CNE draft permits classroom use of generative AI under teacher supervision, requires clear labeling of AI-produced materials, allows objective test scoring and planning support, and blocks automated grading of open-ended work without human review.
• The proposal also vetoes commercial use of educational data and invasive surveillance in schools, and recommends creating a National Observatory and a National AI in Education Program to guide adoption and finance training and infrastructure.
• After the March 16 commission decision, the text will go to public consultation, then to a full council vote not expected before April, and would only take effect after homologation by the Education Ministry.
• Bahia’s education department issued operational guidelines aligned with Brazil’s child-protection and data-privacy laws and opened 500 online training slots for educators for March 3–31 via the Instituto Anísio Teixeira.
• With about 70% of Brazilian high-school internet users already turning to generative tools but only 32% receiving institutional guidance, new research from Rio schools reports creativity setbacks when AI is withdrawn, while a market-moving Citrini report renewed discussion of taxing AI gains to offset job losses.