Draft guidelines on the classification of high-risk AI systems
European Commission (EC)The European Commission (EC) has published for consultation three draft Guidelines on the classification of high-risk artificial intelligence (AI) systems under the AI Act. The objective is to facilitate the uniform application of Article 6 by clarifying when an AI system should be considered high-risk because it is linked to products or safety components subject to European Union harmonization legislation, or because it falls within one of the sensitive use cases listed in Annex III.
Draft guidelines on the classification of high-risk AI systems
Executive Summary
The Guidelines develop the two classification routes set out in Article 6 of the Artificial Intelligence Act: the route under Article 6(1), applicable to AI systems that are products or safety components of products covered by European Union harmonization legislation and subject to third-party conformity assessment; and the route under Article 6(2), applicable to systems whose intended purpose falls within one of the use cases listed in Annex III. In addition, the Guidelines clarify key cross-cutting criteria for classification, including the definition of an AI system, intended purpose, provider documentation, human involvement, complex or agentic systems, and the filter mechanism under Article 6(3).
The public consultation will remain open until 23 June 2026. After reviewing the responses received, the European Commission is expected to adopt the final Guidelines. The drafts also refer to the revised timeline proposed under the AI Omnibus package, which provides for the application of the high-risk rules under Article 6(2) from 2 December 2027, and those under Article 6(1) from 2 August 2028, subject to the final legislative outcome.
Main Content
The draft Guidelines are structured around three main documents:
- Document 1. General principles. Clarifies the common criteria applicable to the high-risk assessment, including the definition of an AI system, the role of intended purpose, provider documentation and the self-assessment of classification.
- Document 2. Annex I. Develops the classification route under Article 6(1), applicable to AI systems that are products or safety components of products covered by European Union harmonization legislation and subject to third-party conformity assessment. It also clarifies that not every AI system embedded in a regulated product is automatically classified as high-risk.
- Document 3. Annex III. Develops the classification route under Article 6(2), applicable to the high-risk use cases listed in Annex III, including biometrics, critical infrastructure, education, employment, essential services, law enforcement, migration, justice and democratic processes. It also explains the filter mechanism under Article 6(3), which allows certain Annex III systems to be excluded from high-risk classification under strict conditions, except in cases involving profiling or complex configurations that materially influence high-risk decisions.
Download the technical note on the Draft guidelines on the classification of high-risk AI systems.