Signal EU AI Act Update: The European Commission Publishes Draft Guidelines on HRAIs

Summary

On 19 May 2026 the European Commission published draft guidelines on classifying high-risk AI systems under the EU AI Act and opened a public consultation that ran until 23 June 2026. The package comprises three documents: general classification principles, guidance on regulated products under Annex I, and a 148-page guidance on high-risk use cases under Annex III, covering areas such as biometrics, education, employment, essential services and law enforcement. The framework places an AI system's intended purpose at the centre of classification, with regulators examining technical documentation, instructions for use, marketing materials and product positioning. The guidance also states that even a component performing only a narrow procedural task within a complex or agentic AI system may still be classified as high-risk if it materially influences an Annex III outcome. Publication came later than planned, since guidance had originally been expected by 2 February 2026. Under the related Digital Omnibus political agreement, the high-risk obligations themselves are now due to take effect on 2 December 2027 for stand-alone Annex III systems and 2 August 2028 for Annex I systems embedded in regulated products.