EU Commission opens public discusson on draft guidelines of classification of high-risk AI systems

The European Commission has opened a public consultation on the draft guidelines of the classification of high-risk artificial intelligence (AI) systems under the EU AI Act. Stakeholders have until 23 June 2026 to provide feedback on the draft guidelines, which are available at https://digital-strategy.ec.europa.eu/en/library/draft-commission-guidelines-classification-high-risk-ai-systems.

Background

The AI Act identifies two categories of high-risk AI systems: AI systems embedded in products regulated under the EU's harmonised legislation on product safety; and AI systems that can significantly affect people's health, safety or fundamental rights in specific use cases listed in the AI Act. 

The draft guidelines are intended to support providers, deployers and other relevant actors in determining whether an AI system falls within these high-risk categories, offer clarifications on the relevant provisions of the AI Act and include practical examples illustrating how the classification should be assessed across different areas and use cases. 

Key content of the draft guidelines

The draft guidelines address the following three main areas:

• General principles for classification. The guidelines set out overarching principles for high-risk classification, including confirming that the system must qualify as an AI system and that classification depends on the intended purpose of the AI system. 
   
• High-Risk classification under Article 6(1) and Annex I. This section covers AI systems that are products or the safety components of products, covered by EU harmonisation legislation. It addresses the classification rationale, the main elements for classification as high-risk under Article 6(1), and the requirement for third-party conformity assessment. The guidelines also discuss mechanisms for minimising the compliance burden for economic operators. 
   
• High-Risk classification under Article 6(2) and Annex III. This section addresses AI systems deployed in specific high-risk use cases and discusses a number of issues, including the role of human involvement in high-risk classification, limitations in some use cases to natural persons, the treatment of AI systems as part of complex systems, products, and services, and the meaning of key phrases such as "intended to be used", "on behalf of," and "in so far as their use is permitted under relevant Union or national law". The guidelines also address the "filter" mechanism that may exempt certain AI systems from high-risk classification, the conditions for that filter to apply, the exception for profiling, and the self-assessment, registration, and monitoring obligations for systems benefiting from the filter. The draft guidelines also provide detailed guidance on high-risk AI systems across the following sectors and use cases listed in Annex III: (i) biometrics, (ii) critical infrastructure, (iii) education and vocational training, (iv) employment, (v) essential private and public services, (vi) law enforcement, (vii) migration, asylum, and border control management, and (viii) administration of justice and democratic processes.

Stakeholder consultation and practical implications

The draft guidelines create the following two priorities:

• The guidelines provide the most detailed interpretive guidance to date on the scope and boundaries of high-risk classification under the AI Act, and stakeholder feedback submitted during the consultation period may shape the final version of these guidelines. Anyone with an interest in the development, deployment, supervision, or use of AI systems (including AI providers and developers, businesses, public authorities, academia, research institutions, civil society organisations, supervisory bodies, and members of the public) is encouraged to review the draft guidelines and engage in the consultation process. 
   
• The provisional agreement between EU institutions on the AI Act (“Digital Omnibus on AI”) delays the compliance deadlines: high-risk rules begin to apply to standalone AI systems from 2 December 2027 while high-risk rules apply to AI built into products will apply from 2 August 2028. In the meantime, organisations developing or deploying AI systems in the EU should review the draft guidelines carefully to assess whether their AI systems may fall within the high-risk classification. In particular, organisations should consider classifying or re-classifying their AI systems in light of the detailed criteria and practical examples set out in the draft guidelines, including systems that may have been previously assessed under earlier interpretive frameworks. For AI systems already classified as high-risk, a reassessment against the guidelines is recommended to confirm that the original classification remains appropriate and identify any changes in scope or categorisation.

For more information on these guidelines and EU regulations in the EU and Hungary, contact your CMS client partner or the CMS experts who contributed to this article.