Commission Issues Draft Guidance on High-Risk AI Classification

The European Commission has published draft guidelines clarifying how to classify artificial intelligence systems as high-risk under the EU Artificial Intelligence Act. The guidelines explain the legal criteria and include practical examples to support providers and deployers in assessing whether their systems fall within the high-risk category. While not legally binding, they reflect the Commission’s interpretation and are expected to guide supervisory authorities in enforcement.

The draft builds on feedback from a public consultation involving industry, civil society, and other stakeholders, as well as input from Member States via the AI Board. The Commission presents the guidelines in a structured and accessible format on the AI Act Single Information Platform. A non-exhaustive list of use-case examples aims to cover a broad range of sectors and is expected to evolve as AI technologies and practices develop.

The guidelines focus on AI systems considered high-risk because of their potential impact on health, safety, or fundamental rights. Providers and deployers of such systems must comply with strict requirements, including risk management, data governance, and human oversight. The document is intended to help legal and technical teams make early and consistent classifications, reducing uncertainty ahead of enforcement.

Following the political agreement on the AI Omnibus, the enforcement timeline has been adjusted. Obligations for high-risk systems used in areas such as biometrics, critical infrastructure, education, employment, migration, asylum, and border control will apply from 2 December 2027. For AI systems embedded in regulated products, including robotics and industrial machinery, the rules will apply from 2 August 2028. A targeted stakeholder consultation on the draft guidelines is open until 23 June 2026, with feedback to be reflected in the final version.