EU AI Act · use case
EU AI Act for EdTech: Admissions, Grading and Exam Proctoring AI
AI that decides admissions, grades students, assigns learning levels or monitors exams is high-risk under Annex III(3) of the EU AI Act. Emotion recognition in education is banned.
Classify my system in 3 minutesFree, no account. Pre-filled for this use case.
Risk tier
High-risk
High-risk (Annex III, point 3); partly prohibited (Art. 5(1)(f))
When it applies
Prohibitions apply since 2 February 2025. High-risk obligations apply from 2 December 2027.
Regulation (EU) 2024/1689, Annex III(3)(a)-(d); Art. 5(1)(f); Art. 26
Education is one of the eight Annex III areas. The EU AI Act lists four educational uses as high-risk: deciding access or admission, evaluating learning outcomes (including when used to steer the learning process), assessing the appropriate level of education a person should receive, and monitoring and detecting prohibited behaviour during tests. Adaptive-learning engines, automated essay scoring, admissions ranking and online proctoring all map onto these entries.
Emotion recognition in educational institutions is prohibited outright, which affects proctoring products that infer stress, attention or “cheating intent” from facial expressions or voice.
Classification
Why this classification applies
Annex III, point 3 targets decisions that shape a person’s educational and professional course. A model that scores an essay and that score feeds the grade is “evaluating learning outcomes”. A recommender that decides which track or level a student is placed in is “assessing the appropriate level”. A proctoring tool flagging suspicious behaviour is “monitoring prohibited behaviour during tests”.
Content tools without evaluative effect, such as an AI that generates practice questions or explains a concept on request, are not listed and are typically minimal-risk, subject to Art. 50 if they interact with students as a chatbot.
Article 5(1)(f) prohibits inferring emotions in educational institutions except for medical or safety reasons. Attention or engagement scores derived from webcam analysis in a classroom or during an online exam are within the ban.
Obligations
What you have to do
- Remove or disable emotion inference features in any product used in education (Art. 5).
- Provider duties for high-risk systems: risk management with attention to bias against students with disabilities, non-native speakers and minority groups; data governance; technical documentation; logging; human oversight; accuracy; conformity assessment; CE marking; EU database registration.
- Human oversight that lets teachers see the basis of a score or flag, override it, and prevent automatic consequences (Art. 14).
- Deployers (schools, universities, exam boards) must inform students and, where relevant, parents that a high-risk system is used, ensure trained oversight, and keep logs (Art. 26). Public-sector deployers must complete a fundamental rights impact assessment (Art. 27).
- Students subject to a decision can ask for an explanation of the AI’s role (Art. 86).
Paperwork
Documents to have on file
Get these documents drafted for your system
Run the free assessment, then unlock the Compliance Pack: a PDF report plus editable first drafts of every required document and a 90-day plan. €49 one-time, no subscription.
Common mistakes
Where companies get this wrong
- Proctoring products that flag “suspicious emotion” or attention drops. This is a prohibited practice, not a configurable feature.
- Automated scoring validated on native speakers only. Data governance requires representativeness for the students actually assessed.
- Universities treating vendor-provided tools as the vendor’s problem. Deployer duties, including student information and oversight, sit with the institution.
- Adaptive platforms that silently decide a student’s level. Assessing the appropriate level is an Annex III entry even when no human ever sees a “decision”.
FAQ
Frequently asked questions
Is an AI tutor chatbot high-risk?
A tutor that explains and answers questions is limited-risk under Article 50 (it must disclose it is AI). It becomes high-risk if its outputs are used to grade, place or admit students.
Can proctoring software use the webcam at all?
Yes, for identity verification and for detecting objectively prohibited behaviour such as a second person in the room. What it must not do is infer emotions or mental states from biometric data.
Do these rules apply to corporate training platforms?
Annex III(3) refers to education and vocational training at all levels, so professional certification and vocational assessment are in scope. Internal corporate learning without evaluative consequences is generally not.
This page is general information about Regulation (EU) 2024/1689, updated 2026-09-17. It is not legal advice; classifications depend on the exact intended purpose of a system. Deadlines reflect the Digital Omnibus adopted in June 2026.