Last updated:
The mandatory process of assigning one of four risk levels (Unacceptable, High, Limited, Minimal) to an AI system under the EU AI Act.
Classification is the gateway to EU compliance. It is determined by the system's intended purpose and the context of its deployment, not its underlying architecture. Getting this wrong leads to massive liability; under-classifying a high-risk system (like an HR ranker) can result in fines up to 7% of turnover, while over-classifying minimal risk tools creates unnecessary bureaucratic overhead.
Real world example:
A developer building an AI-powered Internal Search tool performs a risk classification; they determine it is Minimal Risk because it doesn't fall under EU AI Act Annex III (sensitive sectors) or influence consequential human decisions.
