Annex III: High-Risk AI System Areas
Annex III lists eight categories of AI systems that are classified as high-risk under Article 6(2) when intended for the regulated purposes. The categories cover: biometrics, critical infrastructure, education and vocational training, employment and worker management, essential private and public services, law enforcement, migration and border control, and administration of justice and democratic processes. Each category contains detailed sub-use cases in the official text. Subject to the Article 6(3) carve-out and the Commission's power to amend via delegated acts under Article 7. Always verify sub-bullets on EUR-Lex.
Who does this apply to?
- -Providers determining high-risk classification before placing AI on the EU market
- -Deployers assessing whether their intended use falls under Article 6(2) via Annex III
- -Product, legal, and procurement teams screening AI tools against the eight categories
- -Public-sector bodies deploying AI in law enforcement, migration, justice, or social services
Scenarios
An AI grades university exams for admissions decisions.
AI schedules maintenance shifts for technicians without assessing workers.
The eight Annex III buckets (headline view)
1. Biometric identification and categorisation of natural persons 2. Critical infrastructure management and operation (road traffic, supply of water/gas/heating/electricity, digital infrastructure) 3. Education and vocational training determining access or level 4. Employment, workers management, and access to self-employment (recruitment, selection, performance monitoring) 5. Access to essential private and public services (creditworthiness, emergency services dispatch, insurance pricing, benefits) 6. Law enforcement contexts listed in the annex 7. Migration, asylum, and border control purposes listed 8. Administration of justice and democratic processes (research, interpretation of facts/law, influencing voting)
Always read the exact sub-bullets in EUR-Lex; this summary is not exhaustive.
How to use Annex III in practice
Start from the intended purpose and realistic user-facing functionality. Map each feature to the annex bullets. If multiple bullets hit, assume the strictest obligations apply. Then test Article 6(3) exceptions before concluding you are outside high-risk.
Link to conformity and governance
Once Annex III applies, expect Chapter III design requirements, conformity assessment, CE marking where relevant, Article 49 registration, Article 26/27 deployer duties, and post-market monitoring.
Compliance checklist
- Maintain a living Annex III mapping per product SKU and per geography.
- Archive legal opinions when excluding a borderline use case.
- Synchronise product marketing claims with annex mapping (avoid overselling non-covered uses).
- Feed annex mapping into procurement questionnaires and RFP responses.
- Re-evaluate when you add modalities (voice, video, biometrics) or new customer verticals.
Walk through Annex III with structured questions—free assessment.
Start Free AssessmentRelated Articles
Related annexes
- Annex I — Union harmonisation legislation (Article 6(1) path)
- Annex IV — Technical documentation
Frequently asked questions
Is every recommender system high-risk?
No. Annex III targets specific access and management decisions. Consumer entertainment recommendations rarely fall in scope unless they trigger listed essential-services or employment contexts.
What are the eight high-risk categories listed in Annex III?
Annex III lists: (1) biometrics, (2) critical infrastructure, (3) education and vocational training, (4) employment and workers management, (5) access to essential private and public services, (6) law enforcement, (7) migration, asylum, and border control, and (8) administration of justice and democratic processes.
Can a provider argue their Annex III system is not high-risk?
Yes. Article 6(3) allows a provider to argue that a system listed in Annex III does not pose a significant risk of harm to health, safety, or fundamental rights, considering its intended purpose, output nature, and conditions of use. The provider must document this assessment and notify the relevant market surveillance authority before placing the system on the market.