The European Commission published draft guidelines on the classification of high-risk AI systems under Article 6(5) of the EU AI Act (Regulation (EU) 2024/1689). The guidelines are at consultation stage. The Commission invites feedback from providers, deployers, public authorities, academia, and research institutions by 23 June 2026. The Commission will consider responses before finalising the guidelines.
Article 6 sets two routes to high-risk classification. Under Article 6(1), an AI system is high-risk if it is a safety component of a product covered by EU harmonisation legislation listed in Annex I and the product requires third-party conformity assessment. Under Article 6(2), an AI system is high-risk if it falls within a use case listed in Annex III, covering biometrics, critical infrastructure, education, employment, essential services, law enforcement, migration management, and administration of justice. Article 6(5) allows providers to rebut the Annex III classification by demonstrating the system poses no significant risk of harm. The draft guidelines clarify these classification concepts and provide practical examples of systems that should or should not be classified as high-risk.
Providers must classify their AI systems before the mandatory application dates: 2 December 2027 for stand-alone systems and 2 August 2028 for systems embedded in regulated products. Classification as high-risk triggers obligations including conformity assessment, EU AI database registration, technical documentation, and post-market monitoring under Chapter III of the AI Act. Deployers in regulated sectors (banks, public authorities, insurers, law enforcement agencies) face additional transparency and human oversight obligations.
The final guidelines will not have binding legal force but will represent the Commission's authoritative interpretation and will guide national market surveillance authorities in enforcement. Open questions about how the guidelines interact with the revised application dates agreed in the Omnibus political agreement of 7 May 2026 will be resolved in final published text. The Article 6(5) rebuttal procedure allows providers to argue that an Annex III system poses no significant risk, but the evidentiary standard is not yet settled.
Licentium advises on EU AI Act classification analysis and has a partner network for EU-regulated entities. Work we undertake includes high-risk AI system classification reviews, Article 6(5) rebuttal assessments, Annex III use case mapping, and compliance planning for providers and deployers.