EU AI Act risk categories
Which tier your AI system falls into decides which obligations apply. Each links to the duties it triggers.
Prohibited practice
AI practices banned under Article 5 (e.g. social scoring, certain biom…
High-risk (Annex III use case)
Stand-alone AI whose use case matches one of the eight Annex III domai…
High-risk (Annex I regulated product)
AI that is a safety component of, or is, a product covered by EU harmo…
GPAI model
General-purpose AI model. Provider transparency + documentation + copy…
GPAI model with systemic risk
GPAI model meeting systemic-risk thresholds; additional evaluation, ri…
Limited-risk / transparency
Systems triggering Article 50 transparency duties regardless of high-r…
Minimal risk
Everything else; no mandatory obligations under the Act (voluntary cod…
Stay audit-ready as the Act changes
EU AI Regulation Decoded tracks each obligation and the evidence auditors expect — one practitioner email a week.
Subscribe — free, weeklyNot legal advice. This is a practitioner reference generated from a cited knowledge base built on the primary legal text and official Commission guidance. Verify against the cited primary sources before relying on it for a compliance decision. See editorial standards & methodology.