Conformity assessment and CE marking

The EU AI Act tells providers of high-risk AI systems how to show conformity before a system is placed on the market or put into service. The EU AI Act is Regulation (EU) 2024/1689, as amended by Regulation (EU) 2026/1744. Article 16 makes these steps obligations of the provider:

• Article 16(f). The provider makes sure that the system "undergoes the relevant conformity assessment procedure as referred to in Article 43, prior to its being placed on the market or put into service".
• Article 16(g). The provider prepares an EU declaration of conformity in accordance with Article 47.
• Article 16(h). The provider affixes the CE marking in accordance with Article 48.
• Article 16(i). The provider complies with the registration obligations in Article 49(1).

For all the provider obligations, see Article 16. For the classification of an AI system as high-risk, see Article 6.

Scope and dates

For providers, the EU AI Act excludes Article 6(1) high-risk systems related to Annex I, Section B products from Articles 40–49 (Article 2).

Chapter III sections have different application dates, with different transitional rules for older systems: see Dates (Articles 113 and 111).

Standards and common specifications

For providers of high-risk AI systems and general-purpose AI models, Article 40(1) of the EU AI Act gives a presumption of conformity. It applies when the system or model is in conformity with harmonized standards, or parts of them. The references of harmonized standards must be published in the Official Journal of the European Union in accordance with Regulation (EU) No 1025/2012. The presumption applies to the requirements in Chapter III, Section 2, or, as applicable, to the obligations in Chapter V, Sections 2 and 3. It applies only "to the extent that those standards cover those requirements or obligations".

Article 41 of the EU AI Act gives these rules for common specifications:

• Article 41(1). The Commission is permitted to adopt implementing acts for common specifications only if all the conditions of points (a) and (b) apply.
• Article 41(1), point (a). The Commission asked one or more European standardization organizations for a harmonized standard. Also, one of these conditions applies: no organization accepted the standardization request, the standards were not delivered on time, they "insufficiently address fundamental rights concerns", or they do not comply with the standardization request.
• Article 41(1), point (b). No reference of a harmonized standard for the requirements or obligations is published in the Official Journal, and "no such reference is expected to be published within a reasonable period".
• Article 41(3). Conformity with the common specifications, or parts of them, gives a presumption of conformity "to the extent those common specifications cover those requirements or those obligations".
• Article 41(4). When the reference of a harmonized standard is published in the Official Journal, the Commission repeals the implementing acts, or the parts of them, for the same requirements or obligations.
• Article 41(5). Providers that do not comply with the common specifications must "duly justify that they have adopted technical solutions" that comply with the requirements or obligations "to a level at least equivalent thereto".

Implementing Decision (EU) 2026/2274 cited EN 18286:2026 on 14 October 2026, with presumption only for Annex ZA, Table ZA.1 provisions (Article 17(1) and the first sentence of Article 11(1)), on each row's conditions.

The publication of a reference of a harmonized standard does not show that the standard applies to all the requirements in Chapter III, Section 2. The publication alone also does not make the Annex VI procedure available for a system in Annex III point 1 (see Article 43(1)). For the quality management system in Article 17, see Article 17.

Article 43 — conformity-assessment procedure

Article 43 of the EU AI Act gives the conformity assessment procedure that the provider must use for each type of high-risk AI system. Article 3(20) gives this definition:

‘conformity assessment’ means the process of demonstrating whether the requirements set out in Chapter III, Section 2 relating to a high-risk AI system have been fulfilled;

The provider's conformity assessment procedure depends on the system's Article 6 classification.

High-risk AI system	Conformity assessment procedure	Provision
Annex III point 1 (biometrics), if the provider applied harmonized standards or, where applicable, common specifications	The provider selects Annex VI (internal control) or Annex VII (notified body)	Article 43(1), first subparagraph
Annex III point 1, if one of the conditions (a) to (d) applies	Annex VII	Article 43(1), second subparagraph
Annex III points 2 to 8	Annex VI	Article 43(2)
Annex I, Section A legislation	The procedure of that legislation, with the Chapter III, Section 2 requirements and an assessment of the quality management system	Article 43(3), first subparagraph
Annex I, Section A legislation and an Annex III category	The procedure of the Section A legislation	Article 43(3), last subparagraph
Annex I, Section B legislation, high-risk in Article 6(1)	Article 43 does not apply	Article 2(2)

Some AI systems are in the exclusive competence of the AI Office in Article 75(1), with the scope and exceptions of that paragraph. If such an AI system is high-risk and subject to a third-party conformity assessment in Article 43, the AI Office is responsible for the conformity assessments and tests (Article 75(1e)). These occur before the system is placed on the market or put into service. The Commission entrusts these tests or assessments to notified bodies designated in accordance with the AI Act, which then act on behalf of the Commission. The provider that applied to the Commission for the assessment pays the fees. That provider pays the costs of the entrusted services directly to the notified body.

Article 43(1) — Annex III(1) biometrics

Article 43(1) of the EU AI Act applies to the high-risk AI systems in Annex III point 1. Annex III point 1 gives three categories, "in so far as their use is permitted under relevant Union or national law": remote biometric identification systems, biometric categorization by sensitive or protected attributes or characteristics, and emotion recognition. Article 43(1) applies to all three categories:

For high-risk AI systems listed in point 1 of Annex III, where, in demonstrating the compliance of a high-risk AI system with the requirements set out in Section 2, the provider has applied harmonised standards referred to in Article 40, or, where applicable, common specifications referred to in Article 41, the provider shall opt for one of the following conformity assessment procedures based on:

(a) the internal control referred to in Annex VI; or

(b) the assessment of the quality management system and the assessment of the technical documentation, with the involvement of a notified body, referred to in Annex VII.

In demonstrating the compliance of a high-risk AI system with the requirements set out in Section 2, the provider shall follow the conformity assessment procedure set out in Annex VII where:

(a) harmonised standards referred to in Article 40 do not exist, and common specifications referred to in Article 41 are not available;

(b) the provider has not applied, or has applied only part of, the harmonised standard;

(c) the common specifications referred to in point (a) exist, but the provider has not applied them;

(d) one or more of the harmonised standards referred to in point (a) has been published with a restriction, and only on the part of the standard that was restricted.

For the purposes of the conformity assessment procedure referred to in Annex VII, the provider may choose any of the notified bodies. However, where the high-risk AI system is intended to be put into service by law enforcement, immigration or asylum authorities or by Union institutions, bodies, offices or agencies, the market surveillance authority referred to in Article 74(8) or (9), as applicable, shall act as a notified body.

In condition (d), Annex VII applies only to the restricted part of the standard.

Article 74(9) makes the European Data Protection Supervisor the market surveillance authority for Union institutions, bodies, offices and agencies. This does not apply to the Court of Justice of the European Union in its judicial capacity.

Article 43(2) — other Annex III high-risk systems

Article 43(2) of the EU AI Act gives the procedure for the high-risk AI systems in Annex III points 2 to 8:

For high-risk AI systems referred to in points 2 to 8 of Annex III, providers shall follow the conformity assessment procedure based on internal control as referred to in Annex VI, which does not provide for the involvement of a notified body.

Article 43(2) gives no Annex VII procedure for these systems. Article 43(6) lets the Commission adopt delegated acts that apply Annex VII, or parts of it, to these systems. If a system in these points is also in the Annex I, Section A legislation, Article 43(3) applies.

Article 43(3) — Annex I Section A products

Article 43(3) of the EU AI Act applies to high-risk AI systems in the Union harmonization legislation in Annex I, Section A. Regulation (EU) 2026/1744 replaced Article 43(3) (its Article 1, point (19)). The first subparagraph gives the procedure:

For high-risk AI systems covered by the Union harmonisation legislation listed in Section A of Annex I, the provider of the system shall follow the relevant conformity assessment procedure as required in accordance with the relevant Union harmonisation legislation. The requirements set out in Section 2 of this Chapter shall apply to those high-risk AI systems and shall be part of that assessment. Assessment of the quality management system set out in Article 17 shall also be undertaken, and points 3, 4.3, 4.4. and 4.5, the fifth paragraph of point 4.6 and point 5 of Annex VII shall apply.

The other subparagraphs of Article 43(3) give these rules:

• Notified bodies. Notified bodies that are notified in the Section A legislation "shall have the power to assess the conformity of high-risk AI systems" with the Chapter III, Section 2 requirements. This applies only if the notification procedure in that legislation assessed their compliance with Article 31(4), (5), (10) and (11). Without prejudice to Article 28, these notified bodies must apply for designation in accordance with Chapter III, Section 4 by 28 January 2028.
• Recital (18). In recital (18) of Regulation (EU) 2026/1744, these notified bodies "should have the power to assess the conformity of high-risk AI systems under certain conditions for 18 months from 27 July 2026". The recital tells conformity assessment bodies that want designation that they "can submit an application at any time during and after these 18 months".
• Procedures with no third party. Some Section A legislation lets the manufacturer use a conformity assessment with no third party if the manufacturer applied harmonized standards for all the requirements of that legislation. The manufacturer can use that procedure only if it also applied harmonized standards or, where applicable, Article 41 common specifications for all the requirements in Chapter III, Section 2.
• Selection of procedure. The classification as high-risk in Article 6(1) does not change the selection of procedure that the Section A legislation gives. A high-risk AI system as a safety component of the product does not alone make a third-party procedure necessary, if the Section A legislation does not make it necessary.
• Systems also in Annex III. If a high-risk AI system is in the Section A legislation and in an Annex III category, the provider must use the Section A procedure. The Article 49 registration rules for Annex III systems also apply to that system (see Article 49).

Article 43(4) — substantial modification

Article 43(4) of the EU AI Act gives the rule for a change to a high-risk AI system after its conformity assessment:

High-risk AI systems that have already been subject to a conformity assessment procedure shall undergo a new conformity assessment procedure in the event of a substantial modification, regardless of whether the modified system is intended to be further distributed or continues to be used by the current deployer.

For high-risk AI systems that continue to learn after being placed on the market or put into service, changes to the high-risk AI system and its performance that have been pre-determined by the provider at the moment of the initial conformity assessment and are part of the information contained in the technical documentation referred to in point 2(f) of Annex IV, shall not constitute a substantial modification.

Article 3(23) gives this definition:

‘substantial modification’ means a change to an AI system after its placing on the market or putting into service which is not foreseen or planned in the initial conformity assessment carried out by the provider and as a result of which the compliance of the AI system with the requirements set out in Chapter III, Section 2 is affected or results in a modification to the intended purpose for which the AI system has been assessed;

A change is a substantial modification only if the Article 3(23) definition applies to it. For a system that continues to learn, the pre-determined changes in Annex IV point 2(f) are not a substantial modification. A change that is not one of these pre-determined changes is a substantial modification only if the Article 3(23) definition also applies to it.

These rules are related to Article 43(4):

• Technical documentation. Article 11(1) tells the provider to keep the technical documentation up to date. This obligation is different from Article 43(4). See Articles 8–15.
• Notified body. In the Annex VII procedure, the notified body that issued the certificate assesses each change that "could affect the compliance of the AI system with the requirements or its intended purpose". See Annex VII.
• New provider. In Article 25(1)(b), a distributor, importer, deployer or other third party becomes the provider if it makes a substantial modification to a high-risk AI system and the system stays high-risk. See Article 25.
• Older systems. Article 111(2) uses a different test, "significant changes in their designs", for high-risk AI systems placed on the market or put into service before the Chapter III date. See Dates.

Annex VI — internal control

Annex VI of the EU AI Act gives the conformity assessment procedure based on internal control. In this procedure, the provider does three checks (Annex VI points 2 to 4):

• Point 2. The provider verifies that the established quality management system complies with Article 17.
• Point 3. The provider examines the information in the technical documentation. With it, the provider assesses the compliance of the AI system with the "relevant essential requirements" in Chapter III, Section 2.
• Point 4. The provider verifies that "the design and development process of the AI system and its post-market monitoring as referred to in Article 72 is consistent with the technical documentation".

No notified body takes part (Article 43(2)). The EU declaration of conformity, the CE marking and the registration are other provider obligations. They are in Article 16(g) to (i) and Articles 47 to 49. They are not parts of Annex VI.

For the content that Annex VI checks, see these sections:

• The quality management system: Article 17.
• The technical documentation in Article 11 and Annex IV: Articles 8–15.
• The post-market monitoring system and plan: Article 72.

Annex VII — notified-body assessment

Annex VII of the EU AI Act gives a conformity assessment procedure with two assessments: the quality management system and the technical documentation. A notified body does these assessments. Article 3(22) gives this definition:

‘notified body’ means a conformity assessment body notified in accordance with this Regulation and other relevant Union harmonisation legislation;

A notifying authority notifies a conformity assessment body. The application of the body includes an accreditation certificate "where one exists" (Article 29(2)). If the body cannot give an accreditation certificate, it gives documentary evidence of its compliance with Article 31 (Articles 29(3) and 30(3) to (4)).

Quality management system (Annex VII point 3)

The provider applies to a notified body for the assessment of its quality management system. The application includes these items (point 3.1):

• the name and address of the provider and, if an authorized representative lodges the application, also the name and address of the representative
• the AI systems that the same quality management system applies to
• the technical documentation for each of these AI systems
• the documentation of the quality management system, for all the aspects in Article 17
• a description of the procedures that keep the quality management system "adequate and effective"
• a written declaration that the provider did not lodge the same application with a different notified body.

The notified body assesses if the quality management system satisfies Article 17. It notifies its decision to the provider or its authorized representative, with the conclusions of the assessment and the reasoned assessment decision (point 3.2). The provider continues to keep the approved quality management system "implemented and maintained by the provider so that it remains adequate and efficient" (point 3.3).

Technical documentation (Annex VII point 4)

The provider also applies to a notified body of its selection for the assessment of the technical documentation. This documentation relates to the AI system that the provider intends to place on the market or put into service, and that the quality management system applies to (point 4.1). The application includes these items (point 4.2):

• the name and address of the provider
• the same written declaration as in point 3.1, for this application
• the technical documentation in Annex IV.

The notified body examines the technical documentation. These rules apply when it examines the documentation:

• Data sets (point 4.3). The notified body gets full access to the training, validation and testing data sets, "Where relevant, and limited to what is necessary to fulfil its tasks". This access can be remote, for example through an API, "where appropriate and subject to security safeguards".
• More evidence and tests (point 4.4). The notified body can tell the provider to give more evidence or to do more tests. If the notified body is not satisfied with the tests of the provider, it does "adequate tests" itself, "as appropriate".
• Models (point 4.5). The notified body also gets access to the "training and trained models of the AI system, including its relevant parameters" only if all these conditions apply:
  • The access is "necessary to assess the conformity of the high-risk AI system with the requirements set out in Chapter III, Section 2".
  • All other reasonable means "to verify conformity" are exhausted and "have proven to be insufficient".
  • The notified body makes a reasoned request.
• Protection of rights (point 4.5). Union law on the protection of intellectual property and trade secrets applies to the access to the models.

The notified body notifies its decision to the provider or its authorized representative, with the conclusions and the reasoned assessment decision (point 4.6). If the AI system is in conformity with Chapter III, Section 2, the notified body issues a Union technical documentation assessment certificate. If the AI system is not in conformity, the notified body refuses the certificate and gives detailed reasons. If the AI system does not comply with the requirement for the data used to train it, "re-training of the AI system will be needed prior to the application for a new conformity assessment".

Changes (Annex VII points 3.4 and 4.7)

Annex VII of the EU AI Act gives two different rules for changes:

• Quality management system (point 3.4). The provider tells the notified body about each intended change to the approved quality management system or the "list of AI systems covered by the latter". The notified body examines the change. It decides if the changed quality management system continues to satisfy the requirements in point 3.2 or if a reassessment is necessary. It notifies its reasoned decision to the provider.
• AI system (point 4.7). The notified body that issued the certificate assesses each change to the AI system that "could affect the compliance of the AI system with the requirements or its intended purpose". The provider tells that notified body about such an intended change. The provider also tells it "if it otherwise becomes aware of the occurrence of such changes".
• Decision on a change (point 4.7). The notified body decides if the change needs a new conformity assessment in Article 43(4) or a supplement to the certificate. If it approves the change, it issues a supplement to the certificate.

Surveillance (Annex VII point 5)

In Annex VII of the EU AI Act, the notified body does surveillance of the approved quality management system. The surveillance makes sure that the provider complies with the terms and conditions of that system (point 5.1). The provider gives the notified body access to the premises where "the design, development, testing of the AI systems is taking place", and gives it all necessary information (point 5.2). The notified body does periodic audits and gives the provider an audit report. In these audits, it can do more tests of the AI systems for which it issued a certificate (point 5.3).

Article 44 — certificates

Article 44 of the EU AI Act gives the rules for the certificates that notified bodies issue in accordance with Annex VII. Annex VII gives two different decisions: the approval of the quality management system (point 3.2) and the Union technical documentation assessment certificate (point 4.6).

These rules apply to the certificates:

• Language (Article 44(1)). The certificate is in a language that "can be easily understood by the relevant authorities in the Member State in which the notified body is established".
• Content (Annex VII point 4.6). The certificate gives the name and address of the provider, "the conclusions of the examination, the conditions (if any) for its validity" and the data to identify the AI system. The certificate and its annexes "shall contain all relevant information to allow the conformity of the AI system to be evaluated, and to allow for control of the AI system while in use, where applicable".
• Period (Article 44(2)). Certificates are "valid for the period they indicate". This period is not more than five years for AI systems in Annex I. For AI systems in Annex III, it is not more than four years. "At the request of the provider", the period can be extended for more periods with the same maximum, "based on a re-assessment in accordance with the applicable conformity assessment procedures". A supplement to a certificate "shall remain valid, provided that the certificate which it supplements is valid".
• Corrective action (Article 44(3)). A notified body can find that an AI system no longer complies with the Chapter III, Section 2 requirements. Then it "shall, taking account of the principle of proportionality, suspend or withdraw the certificate issued or impose restrictions on it". This does not apply if the provider takes "appropriate corrective action" that ensures compliance "within an appropriate deadline set by the notified body". The notified body gives reasons for its decision.
• Appeal (Article 44(3)). An appeal procedure against the decisions of notified bodies must be available, also for the certificates.

Articles 28 to 39 give the rules for notifying authorities and notified bodies.

Article 46 — derogation from conformity assessment

Article 46 of the EU AI Act gives a derogation from Article 43. Usually, the provider completes the conformity assessment before the system is placed on the market or put into service (Article 16(f)). Article 46 gives these rules:

• Authorization (Article 46(1)). "By way of derogation from Article 43 and upon a duly justified request", a market surveillance authority is permitted to authorize the placing on the market or putting into service of "specific high-risk AI systems" in its Member State. The reasons must be "exceptional reasons of public security or the protection of life and health of persons, environmental protection or the protection of key industrial and infrastructural assets".
• Period (Article 46(1)). The authorization is "for a limited period while the necessary conformity assessment procedures are being carried out". The procedures are completed "without undue delay".
• Condition (Article 46(3)). The market surveillance authority issues the authorization only if it concludes that the system complies with Chapter III, Section 2. It informs the Commission and the other Member States of each authorization. This obligation "shall not cover sensitive operational data in relation to the activities of law-enforcement authorities".
• Urgency (Article 46(2)). Law-enforcement authorities or civil protection authorities are permitted to put into service "a specific high-risk AI system" without the authorization. This applies only "In a duly justified situation of urgency for exceptional reasons of public security or in the case of specific, substantial and imminent threat to the life or physical safety of natural persons". Also, the authorization must be "requested during or after the use without undue delay".
• Refusal (Article 46(2)). If the authorization is refused, the use stops immediately, and "all the results and outputs of such use shall be immediately discarded".
• Objections (Article 46(4) to (6)). If no Member State and not the Commission object "within 15 calendar days of receipt of the information referred to in paragraph 3", the authorization is deemed justified. If there are objections, the Commission consults the Member State and the operators, and decides. If the Commission considers the authorization unjustified, the market surveillance authority withdraws it.
• Annex I, Section A (Article 46(7)). For high-risk AI systems related to products in the Annex I, Section A legislation, only the derogations from conformity assessment in that legislation apply.

Article 47 — EU declaration of conformity

Article 47 of the EU AI Act tells the provider of a high-risk AI system to prepare an EU declaration of conformity. These rules apply:

• Format and retention (Article 47(1)). The provider prepares "a written machine readable, physical or electronically signed EU declaration of conformity for each high-risk AI system". The declaration identifies the system. The provider keeps it available to the national competent authorities for 10 years after the system is placed on the market or put into service. A copy "shall be submitted to the relevant national competent authorities upon request".
• Statement and content (Article 47(2)). The declaration states that the system complies with the Chapter III, Section 2 requirements. It contains the information in Annex V.
• Language (Article 47(2)). The declaration is translated into a language that the national competent authorities can easily understand. These are the authorities of the Member States where the system is placed on the market or made available.
• One declaration (Article 47(3)). Other Union harmonization legislation can also make an EU declaration of conformity necessary for the system. Then the provider prepares one declaration for all the Union law that applies to the system. The declaration identifies each Union harmonization act that it relates to.
• Responsibility and updates (Article 47(4)). When it prepares the declaration, the provider assumes responsibility for compliance with the Chapter III, Section 2 requirements. The provider keeps the declaration up to date "as appropriate".

Annex V gives the content of the declaration:

Annex V point	Content
1	The name and type of the AI system, and "any additional unambiguous reference allowing the identification and traceability of the AI system"
2	The name and address of the provider or, where applicable, of its authorized representative
3	A statement that the provider alone is responsible for the issue of the declaration
4	A statement that the AI system is in conformity with the AI Act and, if applicable, with other "relevant Union law" that provides for the declaration
5	If the AI system processes personal data, a statement that it complies with Regulations (EU) 2016/679 and (EU) 2018/1725 and Directive (EU) 2016/680
6	"References to any relevant harmonised standards used or any other common specification in relation to which conformity is declared"
7	Where applicable, the name and identification number of the notified body, a description of the procedure, and the identification of the certificate
8	"The place and date of issue of the declaration", the name and function of the signatory, for or on behalf of whom the signatory signed, and a signature

Article 48 — CE marking

Providers of high-risk AI systems must affix the CE marking in accordance with Article 48 of the EU AI Act (Article 16(h)). Article 3(24) gives this definition:

‘CE marking’ means a marking by which a provider indicates that an AI system is in conformity with the requirements set out in Chapter III, Section 2 and other applicable Union harmonisation legislation providing for its affixing;

Article 48 gives these rules:

• General principles (Article 48(1)). The general principles in Article 30 of Regulation (EC) No 765/2008 apply to the CE marking.
• Digital marking (Article 48(2)). For high-risk AI systems "provided digitally", a digital CE marking must be used. It is used "only if it can easily be accessed via the interface from which that system is accessed or via an easily accessible machine-readable code or other electronic means".
• Position (Article 48(3)). The CE marking is affixed visibly, legibly and indelibly. If that is "not possible or not warranted on account of the nature of the high-risk AI system", it is affixed to the packaging or to the accompanying documentation, "as appropriate".
• Notified body number (Article 48(4)). Where applicable, the identification number of the notified body for the Article 43 procedures follows the CE marking. The notified body affixes the number, or the provider or its authorized representative affixes it on the instructions of the notified body.
• Promotional material (Article 48(4)). The number is also given in all promotional material that tells readers that the system fulfills the requirements for CE marking.
• Other Union law (Article 48(5)). Other Union law can also apply to the system and "also provides for the affixing of the CE marking". Then the CE marking also shows that the system fulfills the requirements of that law.

Article 49 — registration in the EU database

Article 49 of the EU AI Act assigns EU database registration duties to providers, authorized representatives where applicable, and specified deployers (Article 71). Each obligation has its actor, its systems and its time:

Article 49 paragraph	Actor	Systems	Time	Content
1	The provider or, where applicable, the authorized representative	High-risk AI systems in Annex III, with the exception of point 2	Before the system is placed on the market or put into service	Annex VIII, Section A
2	The provider or, where applicable, the authorized representative	AI systems for which the provider concluded in Article 6(3) that they are not high-risk	Before the system is placed on the market or put into service	Annex VIII, Section B
3	Deployers that are public authorities, Union institutions, bodies, offices or agencies, or persons that act on their behalf	High-risk AI systems in Annex III, with the exception of point 2	Before putting into service or using the system	Annex VIII, Section C
5	Not specified	High-risk AI systems in Annex III point 2 (critical infrastructure)	Not specified	Registration at national level

In Article 49(1) and (2), the provider or representative registers itself and the system. In Article 49(3), the deployer registers itself, selects the system and registers its use.

Article 49 does not refer to Annex I. Thus, a classification as high-risk in Article 6(1) alone does not cause an Article 49 registration. A system in the Annex I, Section A legislation can also be in an Annex III category. The provider must use the Section A procedure for that system (Article 43(3)). The Article 49 rules for Annex III systems apply to it. The Annex I legislation can have its own registration or notification rules.

Restricted registration (Article 49(4))

Article 49(4) of the EU AI Act applies to the high-risk AI systems in Annex III points 1, 6 and 7, in the areas of law enforcement, migration, asylum and border control management. For these systems, the registration in Article 49(1), (2) and (3) is in a "secure non-public section" of the EU database. It is a registration with less information, not an exception from registration. It includes only this information, as applicable:

(a) Section A, points 1 to 10, of Annex VIII, with the exception of points 6, 8 and 9;

(b) Section B, points 1 to 5, and points 8 and 9 of Annex VIII;

(c) Section C, points 1 to 3, of Annex VIII;

(d) points 1, 2, 3 and 5, of Annex IX.

Annex IX gives the information for testing in real-world conditions in accordance with Article 60. Only the Commission and the national authorities in Article 74(8) have access to these restricted sections.

Annex VIII registration information

Annex VIII of the EU AI Act gives the information for each registration. In each section, the information is given and then kept up to date.

Annex VIII, Section A point (providers, Article 49(1))	Information
1	The name, address and contact details of the provider
2	If a different person submits the information on behalf of the provider, the name, address and contact details of that person
3	The name, address and contact details of the authorized representative, where applicable
4	"The AI system trade name and any additional unambiguous reference allowing the identification and traceability of the AI system"
5	A description of the intended purpose of the AI system and of the components and functions that it supports
6	"A basic and concise description of the information used by the system (data, inputs) and its operating logic"
7	"The status of the AI system (on the market, or in service; no longer placed on the market/in service, recalled)"
8	The type, number and expiry date of the certificate of the notified body, and the name or identification number of that notified body, where applicable
9	A scanned copy of that certificate, where applicable
10	The Member States in which the AI system is placed on the market, put into service or made available
11	A copy of the EU declaration of conformity in Article 47
12	Electronic instructions for use, with the exception of high-risk AI systems "in the areas of law enforcement or migration, asylum and border control management referred to in Annex III, points 1, 6 and 7"
13	A URL for more information (optional)

Annex VIII, Section B point (providers, Article 49(2))	Information
1 to 4	The same information as Section A, points 1 to 4
5	A description of the intended purpose of the AI system
6	The condition or conditions in Article 6(3) on which the AI system is considered to be not high-risk
8	"The status of the AI system (on the market, or in service; no longer placed on the market/in service, recalled)"

Regulation (EU) 2026/1744 deleted Section B, points 7 and 9 (its Article 1, point (42)).

Annex VIII, Section C point (deployers, Article 49(3))	Information
1	The name, address and contact details of the deployer
2	The name, address and contact details of the person that submits the information on behalf of the deployer
3	The URL of the entry of the AI system in the EU database by its provider
4	A summary of the findings of the fundamental rights impact assessment in Article 27
5	A summary of the data protection impact assessment in Article 35 of Regulation (EU) 2016/679 or Article 27 of Directive (EU) 2016/680 "as specified in Article 26(8) of this Regulation, where applicable"

Related registration rules

These EU AI Act rules on other pages are related to Article 49:

• Article 6(3) and 6(4). A provider that considers an Annex III system not high-risk documents its assessment before the system is placed on the market or put into service. See Article 6.
• Article 26(8). Deployers that are public authorities, or Union institutions, bodies, offices or agencies, must not use a high-risk AI system that is not registered in the EU database. They must inform the provider or the distributor. See Article 26.
• Article 5(2). Since 2 February 2025, the use of real-time remote biometric identification in publicly accessible spaces for law enforcement needs a completed fundamental rights impact assessment and a registration in Article 49. "However, in duly justified cases of urgency, the use of such systems may be commenced without the registration in the EU database, provided that such registration is completed without undue delay." See Article 5.

Modulos

In Modulos, the EU AI Act Frameworks are OFF-1 at Organization level and MFF-1 at AI System level: see the Requirements and recommended work sequence.

Related information

• High-risk AI systems: Article 6 classification and the provider requirements.
• Roles and responsibilities: provider, deployer, importer, distributor and representative obligations.
• Post-market monitoring: Article 72 monitoring and Article 73 serious-incident reporting.
• Harmonized standards: harmonized standards for the EU AI Act.
• Operationalizing in Modulos: the Requirement table and the sequence of work in Modulos.
• Control overlap: the Controls that the EU AI Act Frameworks share with other Frameworks in Modulos.

Sources

• Regulation (EU) 2024/1689 (AI Act), as amended, text of 27 July 2026:
  • Articles 2(2), 3(20), 3(22) to 3(24), 5(2), 6, 11(1), 16(f) to (i), 25(1) and 26(8).
  • Articles 28 to 31, 40, 41, 43, 44 and 46 to 49.
  • Articles 74(8) and (9), 75(1) and (1e), 111 and 113.
  • Annexes III, IV point 2(f), V, VI, VII, VIII and IX.
• Regulation (EU) 2026/1744 of 8 July 2026 (Digital Omnibus on AI), Official Journal of 24 July 2026: Article 1, points (2), (19), (31), (39), (40) and (42), Article 4, and recital (18).
• Commission Implementing Decision (EU) 2026/2274 of 12 October 2026, Official Journal of 14 October 2026: Articles 1 and 2 and Annex.

Disclaimer

This page gives general information. It is not legal advice.