EU AI Act risk class checker

Ten questions to work out where your AI system probably sits under the EU AI Act, and which obligations follow. It gives you a direction and the reasoning behind it — the binding answer comes from the text of the Regulation and a lawyer who has seen your system.

Does the system do any of these?
Where you probably sit
Likely minimal risk

Nothing here points at the high-risk or transparency categories. The AI literacy duty under Article 4 still applies to you as provider or deployer, and it is already in force.

This is a direction, not a determination. Classification depends on facts about your system that a form cannot see, and the Regulation applies in stages. Check the current text before you act on it.

Send me the obligations checklist

Questions

Is this a legal assessment?

No. Classification turns on facts about your specific system and on the current text of Regulation (EU) 2024/1689. This tool narrows the question and tells you what to look at; it does not answer it.

Why does it ask about my role first?

Because the same system carries different duties depending on whether you are its provider, a deployer, an importer or a distributor — and a company is often more than one at once. Everything downstream depends on that answer.

What if I only use someone else's AI?

You are likely a deployer, which carries lighter but real duties — including AI literacy for the people operating the system. Fine-tuning or substantially modifying a model can move you into the provider role.

Dates and obligations under the Regulation apply in stages. Verify them against the current text of Regulation (EU) 2024/1689 before acting. We work on this in classification and gap analysis.