Docs

Annex III high-risk AI systems: what falls in scope and how to validate classification

As of 18 November 2026 · checked against Regulation (EU) 2024/1689 and Regulation (EU) 2026/1744

Information only, not legal advice. Classification under Annex III has direct legal consequences. Confirm any concrete decision with qualified counsel and the competent national authority.

Annex III of Regulation (EU) 2024/1689 is the operative list that determines whether an AI system is high-risk by virtue of its intended purpose. If your system is listed, you inherit the full Chapter III obligations: risk management, data governance, technical documentation, logging, human oversight, conformity assessment, post-market monitoring. The obligations apply to Annex III systems from 2 December 2027. The Digital Omnibus moved that date from 2 August 2026, see the deadline timeline. This page maps the eight Annex III areas, walks through the Article 6(3) derogation that lets some providers escape, and shows how to document the classification decision.

The eight Annex III areas

AreaExamplesTypical builders affected
1. BiometricsRemote identification, emotion recognition, biometric categorisationIDV vendors, access control, retail analytics
2. Critical infrastructureSafety components in the management and operation of critical digital infrastructure, road traffic, and the supply of water, gas, heating or electricitySCADA vendors, OT integrators
3. Education and vocational trainingAdmission scoring, assessment, monitoring of prohibited behaviourEdTech, proctoring, LMS analytics
4. Employment and HRRecruitment screening, promotion, termination, task allocationATS vendors, performance tooling, gig dispatch
5. Access to essential servicesCredit scoring, risk assessment and pricing in life and health insurance, public benefits eligibility, emergency call triageFinTech, InsurTech, govtech
6. Law enforcementProfiling, evidence reliability, risk assessment of natural personsPublic-sector vendors
7. Migration and border controlRisk assessment of persons entering, polygraph-type tools, examination of asylum, visa and residence permit applications (verification of travel documents is excluded)Public-sector vendors
8. Administration of justice and democratic processesAssisting a judicial authority in researching and interpreting facts and law, influencing the outcome of elections or voting behaviourLegaltech for courts, political campaign tooling

Article 6(3): the narrow derogation

Even if the intended purpose maps to Annex III, the system is not classified as high-risk if it does not pose a significant risk of harm. Article 6(3) lists four alternative conditions. Meeting any one of them is enough:

An Annex III system that performs profiling of natural persons is always high-risk, whatever the conditions say. A provider that invokes Article 6(3) must document the assessment before placing the system on the market, hand the documentation to the authorities on request (Article 6(4)), and register the system (Article 49(2)).

Validating the classification: a working procedure

A defensible classification process produces three artefacts: a written decision, a register entry, and a review trigger. The procedure below is our suggestion. The Regulation does not prescribe one.

What the EU database expects

Article 49 requires providers of high-risk Annex III systems, and deployers that are public authorities, to register in the EU database the Commission maintains under Article 71. Systems under Annex III, point 2 are registered at national level instead (Article 49(5)). Annex VIII, Section A lists what providers submit, among others:

Common misclassifications we see

Sources

→ See live Annex III updates: https://aiactradar.com

→ Subscribe to webhook: https://docs.aiactradar.com/quickstart