EU AI Act FAQ
How do you classify an AI system under the EU AI Act?
Quick answer
Classify the intended use, not just the product name. Check prohibited practices, the Article 6 product-safety route and Annex III use cases, including relevant exceptions. Then assess any Article 50 transparency duties and the organisation's role. A lower-risk classification does not remove GDPR or other applicable law.
Check both high-risk routes
Article 6 covers certain AI products or safety components under Annex I legislation where the specified third-party conformity conditions apply. Annex III separately lists particular uses in areas such as employment, education and essential services. The Article 6(3) exception requires a careful assessment of whether an Annex III system presents a significant risk, including whether it materially influences a decision. Profiling of natural persons within an Annex III use remains high-risk. A vendor label such as assistant is not a classification decision.
Record the reasoning
Describe the task, the affected people and the decision the output informs. Identify the legal provision and any claimed exception, with supporting supplier information. Transparency duties may apply alongside another classification. Review the record if the intended purpose changes, the system is substantially modified or a new feature affects people differently. Distinguish the classification from the date on which a particular obligation applies.
Reviewed 4 October 2026. Sources: Commission AI Act overview and timetable, classification, deployer duties, literacy guidance and AI Office of Ireland.
Acuity AI helps Irish organisations classify their AI systems accurately under the EU AI Act framework. See our EU AI Act compliance services.