Guideline
EU Guidelines for High-Risk AI Systems
The Commission has published draft guidelines for high-risk AI systems based on stakeholder feedback. They help providers and users to classify AI systems and clarify the criteria for high-risk designation. The guidelines are non-binding but support implementation of the AI Act. A targeted stakeholder consultation runs until 23 July 2026.
High-risk AI, classification and risk levelsGermany: Bundesnetzagentur, BSI and national implementationTimeline, implementation and AI Omnibus
Why it matters
Organisations must correctly classify AI systems to meet legal requirements. The guidelines provide practical examples and criteria for high-risk designation.
Metadata
| Publisher | European Commission |
|---|---|
| Legal status | Draft/proposal |
| Document type | Guideline |
| Jurisdiction | EU |
| Topics | Prohibited practices, High-risk AI, Governance |
| Roles | Provider, Deployer, Public authority |
| Discovered | 2026-05-23T09:47:55Z |
| Last checked | 2026-10-08T18:17:56Z |
| Link status | temporary_error |