Classify your AI use case under the EU AI Act

Evaluate in 6 operational criteria whether a system falls under prohibited practices (Art. 5), high-risk obligations (Art. 6 and Annex I / III), documented Annex III exemption (Art. 6(3)), transparency duties (Art. 50) or minimal risk.

All checks run locally in your browser: no answers are stored or sent. This screening structures technical and documentary readiness and does not replace formal legal counsel.

Load a typical scenario:

Use case parameters

Select the options that best describe your system and organizational role. The classification and evidence checklist update immediately.

1. Organizational role (Art. 3 & Art. 25)

Determines whether Provider conformity obligations (Art. 8–21) or Deployer operational duties (Art. 26–27) apply.

2. Prohibited practices screening (Art. 5)

Prohibitions in force since 2 February 2025.

3. Regulated product safety component (Art. 6(1) & Annex I)

Applies to sectors with EU harmonization laws requiring third-party conformity assessment.

4. High-impact operational domain (Art. 6(2) & Annex III)

Check whether the use case falls within one of the eight sensitive areas listed in Annex III.

5. Decision weight, profiling & Art. 6(3) filter

Determines whether an Annex III system triggers full High-Risk obligations or qualifies for the documented Art. 6(3) exemption.

6. Direct user interaction & synthetic media (Art. 50)

Transparency duties apply both to standalone tools and cumulatively to High-Risk systems.

How the EU AI Act risk classification works

Regulation (EU) 2024/1689 (the EU AI Act) regulates AI systems based on their concrete operational use case and your role in the value chain (Provider vs Deployer).

This classifier evaluates the regulation’s decision gates in order: prohibited practices under Article 5, regulated safety components under Annex I, high-impact domains under Annex III (such as HR recruitment, credit scoring, education and essential services), narrow procedural tasks eligible for the documented Article 6(3) exemption, and specific transparency obligations under Article 50.

The Italian version is available atClassificatore livello di rischio EU AI Act. For broader architectural guardrails see Governed AI and Compliance & auditability.

Scope and limitations

  • This is a preliminary technical and governance screening tool for engineering leads, DPOs and compliance teams; it does not constitute formal legal advice.
  • It focuses on AI systems (Provider and Deployer duties) rather than upstream obligations for General-Purpose AI (GPAI) model providers under Articles 51–55.
  • All selections stay in your browser: copy the generated report to attach it to your internal architecture or pre-DPIA notes.