Sector guide
The EU AI Act for education and training
Annex III point 3 names four distinct education uses, covering the whole path from admission to assessment to invigilation. As in the workplace, inferring emotion is prohibited rather than regulated.
What the Act actually names in this sector
These are the provisions that reach education and training, quoted from the Official Journal. Everything below follows from them.
AI systems intended to be used to determine access or admission or to assign natural persons to educational and vocational training institutions at all levels
Annex III, 3(a) →AI systems intended to be used to evaluate learning outcomes, including when those outcomes are used to steer the learning process
Annex III, 3(b) →AI systems intended to be used for monitoring and detecting prohibited behaviour of students during tests
Annex III, 3(d) →
Common use cases, classified
Prohibited
Banned outright. There is no compliance route — the practice has to stop or change.
- Emotion recognition on students
Inferring emotions in education institutions is prohibited, alongside the workplace.
A narrow medical or safety exception exists in the same point.
Article 5(1)(f) →
High-risk
Permitted, but carries the full Chapter III obligations: risk management, data governance, documentation, logging, human oversight, accuracy and cybersecurity.
- Admissions and placement
Point 3(a) names determining access or admission, or assigning people to educational and vocational training institutions at all levels.
Annex III, 3(a) → - Automated grading and learning assessment
Point 3(b) covers evaluating learning outcomes, including where those outcomes steer the learning process.
Annex III, 3(b) → - Exam proctoring
Point 3(d) names monitoring and detecting prohibited student behaviour during tests.
Annex III, 3(d) →
What gets misread here
Over-classification is the quieter failure: treating something as high-risk when the Act carves it out costs real money and never triggers a complaint, so nobody catches it.
“Engagement detection helps students, so it is a legitimate wellbeing feature.”
Inferring emotions in education institutions is prohibited. The provision's only exception is use put in place for medical or safety reasons.
Article 5(1)(f) →“Adaptive learning just personalises content, so nothing attaches.”
Point 3(b) reaches evaluating learning outcomes including where those outcomes steer the learning process — which is what adaptive systems do.
Annex III, 3(b) →
Dates that matter in this sector
- 2 Dec 2027Annex III high-risk obligations apply — Providers of Annex III high-risk systems (including AI used in employment, education, credit scoring, life and health insurance pricing, essential services, law enforcement and migration) must meet the full Chapter III requirements from 2 December 2027. (date changed by the Omnibus)
- 2 Feb 2025Prohibited practices and AI literacy apply — Since 2 February 2025 the first eight prohibited AI practices in Article 5 have been unlawful across the EU, and every provider and deployer has owed an AI literacy duty under Article 4. Two further prohibitions were added later and apply from 2 December 2026.
Dates are as amended by the Digital Omnibus. See the full timeline for what each one covers.
What the Commission has said
These guidelines are not binding, but they are the Commission’s own reading of the provisions above, and they are the first place a supervisory authority will look. Where they and this guide differ, follow them.
- Guidelines on prohibited AI practices ↗European Commission, February 2025 — interprets Article 5
What to do next
Screen your own systems with the scope finder, run a single system through the risk classifier, or work the obligations in order with the checklist.
This guide points you at the provisions that govern your sector. It is not legal advice, and whether a particular system meets a provision turns on facts about that system and on whether you hold it as provider or deployer — see Article 3 and Article 25.