High-risk AI systems under the EU AI Act

There are two independent routes. The first is Article 6(1): the system is intended to be used as a safety component of a product, or is itself a product, covered by the Union harmonisation legislation listed in Annex I, and that product must undergo a third-party conformity assessment before being placed on the market. Both conditions must be met. Annex I lists the familiar product regimes, among them machinery, lifts, toys, radio equipment, medical devices, in vitro diagnostics, motor vehicles and civil aviation.

The second route is Article 6(2): the system is used in one of the areas listed in Annex III, whether or not it forms part of a regulated product. Annex III is written around intended purpose, not technology. What matters is what the system is intended to be used for, as described in the instructions for use and the provider's own materials, rather than the model or the vendor behind it.

Four entries do most of the work in the private sector. Point 4 covers employment: recruitment and selection, including targeted job advertisements, filtering applications and evaluating candidates, plus decisions on promotion, termination, task allocation and performance monitoring. Point 5(b) covers evaluating creditworthiness or establishing a credit score, with fraud detection carved out. Point 5(c) covers risk assessment and pricing in life and health insurance. Point 3 covers education and vocational training, including admission, evaluating learning outcomes and proctoring.

The rest are narrower. Point 1 covers biometrics: remote biometric identification, biometric categorisation by sensitive attributes and emotion recognition, where not already prohibited by Article 5. Point 2 covers AI used as a safety component in critical infrastructure such as road traffic or the supply of water, gas, heating or electricity. Points 6 to 8 cover law enforcement, migration and border control, and the administration of justice and democratic processes, and are addressed mainly to public bodies.

Article 6(3) provides a derogation. An Annex III system is not high-risk where it poses no significant risk of harm to health, safety or fundamental rights, including by not materially influencing the outcome of decision making, and where one of four conditions applies: it performs a narrow procedural task; it improves the result of a previously completed human activity; it detects decision-making patterns or deviations without replacing or influencing an earlier human assessment absent proper review; or it performs a preparatory task.

Two guards sit around it. A system that performs profiling of natural persons is always high-risk, whichever condition it appears to satisfy. And the derogation is not self-service: a provider concluding that an Annex III system is not high-risk must document that assessment before placing the system on the market, register it in the EU database under Article 49(2), and produce the documentation for national competent authorities on request (Article 6(4)).

Chapter III, Section 2 sets the substance: a lifecycle risk management system (Article 9), data governance (Article 10), technical documentation to Annex IV (Article 11), automatic logging (Article 12), instructions for use (Article 13), human oversight (Article 14) and declared accuracy, robustness and cybersecurity (Article 15). Article 16 turns those into operational duties: a quality management system (Article 17), documentation kept for the authorities (Article 18), logs kept where under the provider's control (Article 19), conformity assessment before market placement (Article 43), an EU declaration of conformity (Article 47), CE marking (Article 48), registration (Article 49) and corrective action where a system is not in conformity (Article 20).

The provider develops the system, or has it developed, and places it on the market or puts it into service under its own name or trademark (Article 3(3)); it carries Article 16. The deployer uses the system under its authority in a professional context (Article 3(4)) and carries Article 26: use according to the instructions, human oversight, input data, monitoring, logs, and information to workers and affected persons. Importers verify the conformity assessment, the technical documentation, the CE marking and the declaration (Article 23).

Distributors verify the CE marking, the declaration and the instructions before making a system available, and must act where they have reason to consider it non-conforming (Article 24). Roles can change. Under Article 25(1) a distributor, importer, deployer or other third party becomes the provider, with the Article 16 duties, if it puts its name or trademark on a high-risk system already on the market, substantially modifies one, or changes the intended purpose of a system so that it becomes high-risk.

Regulation (EU) 2026/1744, in force since 27 July 2026, moved both dates. Chapter III, Sections 1 to 3 applies to Annex III systems from 2 December 2027, which covers the provider requirements as well as the deployer duties in Articles 26 and 27. Article 6(1) and the corresponding obligations for Annex I product systems apply from 2 August 2028. Systems already on the market have their own treatment under Article 111(2).