EU AI Act Risk Classification Calculator

How the EU AI Act risk classification works

The Regulation asks four questions in a fixed order, and each answer decides whether the next one matters. This tool walks the same sequence and shows which provision drove each conclusion.

1. Is the system in scope at all? Article 2 catches providers placing an AI system on the EU market wherever they are established, and providers and deployers outside the Union whose system output is used in the EU. Military, national security and purely personal use are excluded.
2. Is it a prohibited practice? Article 5 bans ten practices outright. If one applies, nothing further is assessable: the system cannot be placed on the market or used, and the exposure is up to €35 million or 7% of turnover.
3. Is it high-risk? Two independent routes: a safety component of a product covered by the Annex I harmonisation legislation, or a use case in one of the eight Annex III areas. The routes carry different deadlines: 2 August 2028 and 2 December 2027 respectively.
4. Do transparency duties attach? Article 50 applies regardless of tier where a system interacts with people, generates synthetic content, produces deep fakes, or infers emotion or biometric categories.

What the tool produces

A classification for each tier with the provision that produced it, a list of the obligations that follow for your role, the dates each obligation applies from, and the open questions that need a human answer before the classification can be relied on. The result can be exported as a dated assessment record for your compliance file.

What it does not do

It is not a conformity assessment. A conformity assessment under Article 43 is a formal procedure that in defined cases requires a notified body, and it cannot be completed by answering questions. The tool is a preliminary screen and an evidence-planning aid: useful for scoping the work and for showing your reasoning, not for concluding it.