TLDR : The European Commission must publish guidelines by February 2026 for the application of the AI Act, which regulates high-risk AI systems. A public consultation is open to clarify the classification of these systems and the responsibilities of the actors involved in their development and deployment.
By February 2, 2026, the European Commission will need to publish guidelines on implementing Article 6 of the AI Act, which governs the classification of high-risk AI systems (AIS). To support this reflection, a public consultation was launched on June 6 to clarify the classification criteria, as well as the obligations and responsibilities of the stakeholders involved in the development and deployment of these systems.
Although it came into force on August 1, 2024, the application of the AI Act or AIA will be gradual until August 2027. Requirements related to high-risk AIS are expected to be fully effective in August 2026. Article 6 of the AI Act distinguishes two types of high-risk AIS:
- AIS intended to be used as a safety component of a product covered by Union harmonization legislation mentioned in Annex I of the AI Act or AIS that themselves constitute such a product (art. 6(1));
- AIS referred to in Annex III of the regulation which, due to their purpose, present a significant risk to health, safety, or fundamental rights (art. 6(2)).
All stakeholders, from high-risk system suppliers and developers to public and private users, including academic institutions, research centers, civil society, supervisory authorities, and citizens, are invited to contribute to this consultation which will be open until July 18. It will allow the Commission to gather concrete usage examples and identify the operational needs of sector stakeholders.