EU AI Act Article 6 high-risk classification: Annex I products, Annex III uses and the safety-component rule

Why this page: the site explains who the Act reaches and what a non-EU supplier owes. A supplier’s first practical question is whether a given AI feature is “high-risk” at all. Article 6 answers that. This page quotes the Service Desk text, which the Service Desk says is based on the EUR-Lex consolidated version as at 27 July 2026. It is not legal advice; see the EU AI Act explainer.

Two routes into “high-risk”

The Service Desk’s own summary says an AI system is high-risk if it is intended to be used as a safety component of a product, or is itself a product, covered by the EU legislation listed in Annex I, and that product is required to undergo a third-party conformity assessment, or if it is listed in Annex III. The summaries there are not legal binding, so read the paragraph text below.

Route 1, Annex I products. Article 6(1) requires both of these: the AI system is intended to be used as a safety component of a product, or the AI system is itself a product, covered by the Union harmonisation legislation listed in Annex I, and the product is required to undergo a third-party conformity assessment. Annex I is split into Section A. List of Union harmonisation legislation based on the New Legislative Framework and Section B. List of other Union harmonisation legislation.

Route 2, Annex III uses. Article 6(2) adds that AI systems referred to in Annex III shall be considered to be high-risk. Annex III lists areas such as biometrics and, under critical infrastructure, AI systems intended to be used as safety components in the management and operation of critical digital infrastructure, road traffic, or in the supply of water, gas, heating or electricity.

The new safety-component paragraphs

The Service Desk marks paragraphs 1a to 1c as new from the Digital Omnibus. Paragraph 1a says that AI systems that are solely used for non-safety related aspects of user assistance, performance optimisation, service efficiency, automation or convenience or quality control shall not qualify as safety components. Paragraph 1b follows with AI systems the failure or malfunctioning of which would endanger health and safety shall qualify as safety components.

Paragraph 1c says a product required to undergo a third-party conformity assessment solely due to risks other than risks to health and safety does not meet the condition in paragraph 1, point (b).

Editorial judgment: for a manufacturer, the practical test becomes a failure-effect question: would failure or malfunction of the AI feature endanger health and safety? That is a question an engineering team already answers in a hazard analysis, so the safety case is the natural place to record the answer. Whether a given feature falls inside 1a or 1b is a legal call for counsel.

The Article 6(3) route out of Annex III

Article 6(3) says an Annex III system shall not be considered to be high-risk where it does not pose a significant risk of harm to the health, safety or fundamental rights of natural persons, including by not materially influencing the outcome of decision making. It lists conditions such as that the system is intended to perform a narrow procedural task. It does not apply where the system performs profiling of natural persons: such a system shall always be considered to be high-risk.

A provider relying on the carve-out must document its assessment before that system is placed on the market or put into service and, on request, give that documentation to national authorities.

Dates

The Commission’s AI Act page says the Annex I rules have an extended transition period until 2 August 2028 and the Annex III use cases have been extended to 2 December 2027. Check the current text before relying on a date.

Questions

Does Article 6 decide whether defense or military AI is covered?

No. Scope and exclusions are in Article 2, covered in AI Act scope for defense and aviation. Article 6 only classifies systems that are already in scope.

Where do Annex I Section B products such as motor vehicles fall?

That is a separate question from the Article 6 test above. See Motor-vehicle AI under the AI Act for how the site reads the Section B entry for vehicles.