EU AI Act risk classification
Walk through prohibited practices, high-risk routes (product safety and Annex III), the Article 6(3) derogation, and transparency triggers—one step at a time, with supporting guidance.
Based on Regulation (EU) 2024/1689 (AI Act) classification rules and published Commission guidance themes. Not legal advice. Multi-purpose systems may need more than one run.
How classification works
The AI Act does not use a single “risk score.” Instead, you apply legal tests in order: prohibited practices, then high-risk classification (product/safety route and Annex III use cases, with a narrow Article 6(3) derogation), then remaining transparency duties.
- Base answers on intended purpose and real use, not product marketing names.
- Run the tool separately for each distinct use case if multi-purpose.
- Article 6(3) “not high-risk” claims require documented assessment—and may still need notification.
- Commission guidelines on prohibited practices and high-risk classification provide practical examples.
Supporting information
Examples & references
Legal anchors used in this tool
Article 5
Prohibited AI practices (unacceptable risk). Checked before high-risk analysis.
Article 6 + Annexes I & III
High-risk routes: regulated products/safety components, and standalone Annex III use cases, with Article 6(3) derogation.
Article 50
Transparency obligations for certain interactions and synthetic content—often labelled “limited risk.”