Initial reactions on what providers, deployers and downstream actors should take away from the European Commission’s long-awaited Article 6 Guidelines on the classification of high-risk AI systems.
On 19 May 2026, the European Commission published its long-anticipated Draft Commission Guidelines on the classification of high-risk AI systems under Article 6 of Regulation (EU) 2024/1689 for stakeholder consultation (“draft Guidelines”). Across three documents, general principles, the Annex I (product safety) route and the Annex III (use case) route, the Commission sets out, for the first time at this level of detail, how it reads the single most consequential classification question under the AI Act: when does a system fall inside the high-risk regime, and when does it stay outside?