High-Risk AI Classification Assessment

Document whether potentially high-risk AI systems meet applicable classification criteria and any relied-upon exception.

Use this control to document a reasoned high-risk classification, not to select the least burdensome label.

1. Identify the system's intended purpose, actual use and the organization's role.
2. Assess the relevant Article 6 route, including product-safety criteria and the specified Annex III use cases. Treat Annex I Sections A and B separately when determining which requirements apply.
3. If relying on the Article 6(3) exception for an Annex III system, document the full reasoning, including the rule for profiling natural persons. Arrange the applicable registration before market placement or service.
4. Reassess after material changes or a regulator's challenge, and assign any resulting compliance actions. Keep the assessment and supporting evidence accessible to the responsible reviewer and competent authority where required.

Before market placement or putting into service; after material design or purpose changes; on regulatory challenge or relevant legal change.

Classification record describing the system and intended use Relevant product legislation or Annex III analysis Documented exception reasoning and supporting evidence Registration evidence where required Approvals, regulator correspondence and reassessment actions

Select an assessment concluding that a system is not high-risk. Check that it tests the actual criteria rather than relying on a supplier label. Confirm that material changes, required registration and any authority decision have been addressed.