Risk tiers
How the AI Act sorts systems by risk, and which 'bucket' your tool falls into.
What this topic covers
The AI Act uses a risk-based approach: the greater the potential harm to people, the stricter the duties. This section explains the four risk tiers, how they differ, and how to work out which one a specific AI tool you use belongs to.
- Unacceptable risk — uses prohibited under Article 5.
- High risk — Annex III areas with strict duties even for deployers.
- Limited risk — chatbots and AI content with a transparency duty (Article 50).
- Minimal risk — most everyday AI, barely regulated.
- Classification depends on the specific use case and can change.
This section is an overview and is being expanded over time. It is general information, not legal advice — for a specific case, consult a qualified professional.