The European Commission has opened a targeted consultation on draft guidelines for the classification of high-risk AI systems under Article 6 of Regulation (EU) 2024/1689 (the AI Act), which entered into force on 1 August 2024. The consultation period closes 23 June 2026. The draft is at proposed stage; the Commission will review submissions before publishing a final version intended to support market surveillance authorities and AI system providers in applying the Article 6 classification rules.
Article 6(1) of the AI Act classifies an AI system as high-risk when it is intended as a safety component of a product subject to EU harmonisation legislation listed in Annex I and the product requires third-party conformity assessment under that legislation. Article 6(2) applies the high-risk classification to AI systems falling within the use-case categories in Annex III, which cover biometric identification, critical infrastructure, education and vocational training, employment and worker management, access to essential services, law enforcement, migration and border control, and administration of justice. The draft guidelines provide practical examples for each Annex III category and a step-by-step methodology for providers to assess classification before placing a system on the market.
Providers of AI systems intended for the EU market must determine whether their system qualifies as high-risk before the compliance deadlines in Articles 85 and 110 of the AI Act take effect. Mis-classification generates two categories of risk: placing an unregistered high-risk system on the market triggers infringement proceedings and financial penalties up to EUR 30 million or 3% of global annual turnover under Article 99(3); over-classification imposes disproportionate conformity assessment costs without legal obligation. Deployers in credit, insurance, recruitment, critical infrastructure management, and essential public service sectors should map their vendor AI systems against the Annex III categories before the consultation closes.
The draft guidelines do not address prohibited AI practices under Article 5, nor do they bind the European AI Office or member state market surveillance authorities, which retain independent enforcement authority. The Commission may revise the classification methodology in the final guidelines following consultation feedback received by 23 June 2026. Providers relying on the Article 6(3) exemption for AI systems not falling within the Annex III use cases must document their exemption reasoning and retain that documentation for inspection by national competent authorities.
Licentium advises AI developers, deployers, and regulated-sector firms on EU AI Act compliance, classification analysis, and conformity assessment planning. We may advise on this matter or refer you to our partner network for EU AI regulatory counsel. Work we undertake includes Annex III use-case assessments, technical documentation preparation, fundamental rights impact assessments, and support for registration in the EU AI Act database.