June 23, 2026. That’s the date compliance teams need to mark now.
The European Commission released draft guidelines, in mid-May 2026, for classifying AI systems as high-risk under Article 6 of Regulation (EU) 2024/1689, the official EU AI Act text, and opened a public comment period running through June 23. The guidelines address a question that has sat unresolved for many organizations: does their AI system qualify as high-risk, and under which track?
Two tracks exist. Article 6(1) covers AI systems embedded in products already regulated under EU product safety law, medical devices, machinery, vehicles, where the AI component triggers high-risk classification by virtue of the product category. Article 6(2) covers AI systems that operate in one of eight sensitive areas listed in Annex III of the regulation: biometrics, critical infrastructure, education and vocational training, employment and workforce management, essential services, law enforcement, migration and border control, and administration of justice. If a system touches any of those domains, it’s a candidate for high-risk classification, though the guidelines also address how to assess whether a system poses significant risk within those categories, not just whether it appears in the list.
Why does the timing matter? The high-risk provisions of the EU AI Act take effect August 2, 2026. The Commission initially targeted February 2026 for releasing these guidelines, per reporting from IAPP, and the delay compressed the runway. Organizations now have roughly ten weeks between the guidelines release and the enforcement date. The consultation window, June 23, sits at the midpoint of that runway. After consultation closes, the Commission will finalize the text. That finalization timeline hasn’t been announced. Organizations that want to influence the final language must comment before June 23; those waiting for final guidance before beginning their classification work are accepting meaningful timeline risk.
Timeline
This is the third EU AI Act compliance activation in May 2026. The Article 50 transparency guidelines consultation closed June 3. The Omnibus finalization resolved outstanding definitional questions earlier this month. Now the classification machinery itself is operational.
The real question isn’t whether your AI system is on the Annex III list. It’s whether your legal and product teams have walked through the Article 6(2) significant risk assessment, because Annex III inclusion is a starting trigger, not a final determination.
Don’t expect the final guidelines to be materially different from the draft. Comment periods for technical guidelines rarely produce structural reversals. Organizations that treat the draft as effectively final, while still submitting comments on specific points, are making the right call.
Who This Affects
The two-consultation overlap, June 3 for Article 50 transparency, June 23 for Article 6 high-risk classification, means many compliance teams are managing parallel workstreams right now. Those aren’t the same obligation. An Article 6 high-risk determination drives conformity assessment requirements, technical documentation, and registration obligations. Article 50 transparency requirements apply to a different, broader set of systems. Don’t let the deadline proximity create category confusion.
The catch is the sequencing risk: organizations that misclassify, either wrongly excluding themselves from high-risk status, or failing to complete conformity assessments, will be operating out of compliance on August 2 with no procedural defense. The finalization timeline is unknown. The enforcement date isn’t.