EU AI Act Risk Evaluation
Paste spec • Evaluate • get a deterministic classification
Prohibited (Art. 5) • High-risk (Annex III) • Transparency (Art. 50) • Unknown • minimal follow-ups
Describe your AI system
Deterministic • Follow-ups only if needed
Paste spec → facts (yes/no/unknown) → deterministic evaluator → audit trail. Not legal advice.
Click Evaluate to start. If anything is unknown, we’ll ask the minimum follow-ups.
Learn more: classification logic, FAQ, and guides
What this tool evaluates
- Potential prohibited signals (Article 5)
- Potential high-risk signals (Annex III)
- Potential transparency obligations (Article 50)
- Deterministic output + audit trail + follow-ups only when facts are missing
FAQ (short)
Is this legal advice?
No. This is deterministic decision-support meant to help you triage risk categories and produce reviewable evidence.
Why deterministic?
The classification step is rule-based. Any AI assist is used only to map your spec into explicit yes/no/unknown facts.
What should I include in my spec?
State the deployment context, who the system affects, and what decisions it supports (employment, education, services, biometrics, etc.).
Policy Guides
Based on Regulation (EU) 2024/1689 • No hallucinations in the gate • Reviewable output