What is AI Risk Classification?
Risk classification is deciding which category an AI system falls into, and therefore which obligations apply to it. Under the EU AI Act, some practices are prohibited, the use cases listed in Annex III and certain products are high-risk, and other systems carry lighter or no specific obligations.
Annex III high-risk areas
Annex III lists areas including biometrics, critical infrastructure, education, employment, access to essential services (such as creditworthiness and life and health insurance pricing), law enforcement, migration and the administration of justice. After the 2026 amendment, the high-risk rules for these areas apply from 2 December 2027.
What follows from it
High-risk systems carry obligations including a risk management system (Article 9), data governance (Article 10), technical documentation (Article 11), record-keeping (Article 12), transparency to deployers (Article 13) and human oversight (Article 14); deployers have their own (Article 26).
Classifying with NOMOS
The EU AI Act is published on the Exchange as an authority: describe a system's use and get its classification under the Act's rules, with the article that decided it. NOMOS compiled it from the official text; it has not been issued or attested by the EU.
Example
An insurer uses a model to price health insurance.
The team assumes it is limited risk because the model is "only pricing".
Asked with the facts of the use, the Act's rules classify it as high-risk under Annex III, point 5(c), and the obligations that follow are known from the start.
Frequently asked questions
Related concepts
See it in practice
Upload a policy, confirm its rules, seal them: any system can then check an action against them.