European Commission delivers draft high-risk AI guidelines after delays

Covered entities are seeing progress toward long-sought guidance around high-risk artificial intelligence systems under the EU AI Act.

After various delays, the European Commission released draft guidelines and opened a public consultation 19 May aimed at supporting "providers, deployers and other relevant actors in determining whether an AI system falls within the high-risk category." The three-phased guide brings clarity around implementation of high-risk requirements while offering examples "to illustrate how the classification should be assessed in different areas and use cases."

The first section of the guide covers general principles for classifying high-risk AI while there are two sections that delve into the approach and specific requirements around the two high-risk categories under Article 6 of the act. Article 6(1) and Annex I applies to AI systems used in the context of product safety while Article 6(2) and Annex III cover high-risk systems deployed across eight areas, including biometrics, education, employment and law enforcement.

The Commission stated the examples provided in its guidance "strive to cover all areas and use cases, but they are not to be considered as exhaustive and may be updated over time."

The comment period on the draft guidelines is open through 23 June. It is not clear when the Commission will finalize the guidance or whether it will seek further public comment on subsequent drafts.

The IAPP Resource Center carries a series on AI Act operational impacts that includes deep analysis of obligations on nonproviders and providers of high-risk systems, respectively.

The holdup

The guidelines were initially due 2 Feb. ahead of AI Act provisions on high-risk systems taking effect 2 Aug.

Following a delay on the initial February deadline, a Commission spokesperson told Euractiv further delay was imminent with the draft guidance "currently subject to a revised timeline" to allow for the consideration and incorporation of as much stakeholder feedback as possible.

The Commission's lag played a key role in the expedited path of the recently finalized Digital Omnibus on AI, which was first introduced November 2025. Stakeholders, including more than 110 EU-based businesses, lobbied for a two-year pause on enforcing high-risk rules due in large part to the absence of appropriate guidance and standards in a shrinking grace period.

The provisional agreement between EU institutions on AI Act reforms included delayed compliance deadlines for high-risk rules. The implementation date for rules on stand-alone high-risk systems is now 2 Dec. 2027 and 2 Aug. 2028 for high-risk systems embedded in products.

During a European Parliament Committee on Civil Liberties, Justice and Home Affairs hearing 26 Jan., European Commission Deputy Director-General for Communications Networks, Content and Technology Renate Nikolay indicated the AI omnibus would provide "a bit more time to work on either guidelines or specification or standards, so that we can provide this legal certainty for the sector, for the innovators, so that we have the full system in place."