Express + expert help

View pricing →
Skip to main content
guideAIAct, CRA, CEMark, MachineryReg7 min read

CE marking AI-enabled hardware — CRA + AI Act overlap

When AI in a hardware product is high-risk under the AI Act, the dates set by Regulation (EU) 2026/1744, and how CE marking and the CRA fit together.

By CenitiaUpdated

Fact-checked against the law as of Reviewed by Vladimír Vician on How we check

AI inside a device does not by itself make the device a high-risk AI system. For hardware, the AI Act (Regulation (EU) 2024/1689) asks two questions: is the AI a safety component of a product covered by the product legislation listed in its Annex I, and does that legislation require third-party conformity assessment of the product? This guide sets out that test as amended by Regulation (EU) 2026/1744 (the Digital Omnibus on AI), the application dates as of 22 November 2026, and how the AI Act, the CE marking and the Cyber Resilience Act (CRA, Regulation (EU) 2024/2847) connect.

When embedded AI is high-risk (AI Act Article 6(1))

Under Article 6(1) of Regulation (EU) 2024/1689, an AI system is high-risk where both of the following conditions are fulfilled:

  • (a) it 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
  • (b) that product is required to undergo a third-party conformity assessment under that legislation before it is placed on the market or put into service.

Regulation (EU) 2026/1744 added three clarifications to Article 6 and rewrote one definition:

  • Article 6(1a). AI systems used solely for non-safety related aspects of user assistance, performance optimisation, service efficiency, automation, convenience or quality control do not qualify as safety components.
  • Article 6(1b). AI systems whose failure or malfunctioning would endanger health and safety do qualify as safety components.
  • Article 6(1c). A product that needs third-party conformity assessment solely because of risks other than risks to health and safety — in particular risks relating to radio spectrum or electromagnetic interference that do not affect health and safety — does not fulfil condition (b).
  • Article 3(14). A safety component is a component that fulfils a safety function, meaning its intended purpose is to prevent or mitigate risks to the health and safety of persons or property, or whose failure or malfunctioning endangers them.

For a radio product this matters directly. The Radio Equipment Directive 2014/53/EU is listed in Annex I, but a notified-body procedure that is needed only because of its radio-spectrum requirement does not make embedded AI high-risk (Article 6(1c)).

Annex III is a separate route (Article 6(2)). It lists use cases, among them biometrics. Point 1(a) covers remote biometric identification systems, but not AI systems intended for biometric verification whose sole purpose is to confirm that a specific person is the person he or she claims to be. A feature that only checks a claimed identity is therefore not caught by point 1(a); assess the actual intended function against the other points of Annex III.

Annex I Section A and Section B — where machinery now sits

Annex I has two lists, and the consequences differ.

Section A lists New Legislative Framework acts. After the 2026 amendment it covers toys, recreational craft, lifts, equipment for explosive atmospheres (Directive 2014/34/EU), radio equipment (Directive 2014/53/EU), pressure equipment, cableway installations, personal protective equipment, gas appliances, medical devices (Regulation (EU) 2017/745) and in vitro diagnostic medical devices (Regulation (EU) 2017/746). For a high-risk AI system under these acts, the requirements of Chapter III, Section 2, of the AI Act apply and are assessed as part of the product's own conformity assessment procedure (Article 43(3)).

Section B lists other legislation, such as vehicle and aviation rules. Regulation (EU) 2026/1744 deleted the Machinery Directive 2006/42/EC from Section A and added the Machinery Regulation (EU) 2023/1230 to Section B. For AI systems related to Section B products, only Article 6(1), Article 60a and Articles 102 to 112 of the AI Act apply (Article 2(2)). The AI requirements reach machinery through the Machinery Regulation itself: its amended Article 8 requires the Commission to add health and safety requirements for high-risk AI systems to Annex III of that Regulation by delegated acts, which are to apply by 2 August 2028.

Application dates as of 22 November 2026

Regulation (EU) 2026/1744 was published in the Official Journal on 24 July 2026 and entered into force on 27 July 2026. It replaced the application dates in Article 113 of the AI Act for the high-risk rules (Chapter III, Sections 1 to 3):

RuleDate
Chapters I and II, including the original prohibited practicesApply since 2 February 2025
New prohibitions in Article 5(1), points (ba) and (bb)Apply from 2 December 2026
Article 50(2) marking of synthetic content, for systems placed on the market before 2 August 2026Comply by 2 December 2026 (Article 111(4))
High-risk AI systems under Article 6(2) and Annex IIIApply from 2 December 2027
High-risk AI systems under Article 6(1) and Annex I (AI in products)Apply from 2 August 2028

The postponement concerns the high-risk rules. It is not a reason to delay other AI Act duties, or product-law duties, that already apply.

Under Article 111(2) as amended, high-risk AI systems placed on the market or put into service before the Chapter III application date come under the Regulation only if, from that date, they are subject to significant changes in their designs. Recital 39 of Regulation (EU) 2026/1744 explains that this is read per type and model: once one unit has been lawfully placed on the market before the date, further units of the same unchanged design may follow.

Conformity assessment, declaration and CE marking

Annex I Section A products (Article 43(3)). The provider follows the conformity assessment procedure of the product legislation, and the AI Act requirements are part of that assessment. Two points in the paragraph as replaced by Regulation (EU) 2026/1744 matter for hardware manufacturers:

  • Where the product legislation lets the manufacturer avoid third-party assessment by applying harmonised standards, that option remains available only if harmonised standards, or common specifications, covering all the Chapter III, Section 2, requirements have also been applied.
  • Manufacturers are not required to choose a third-party procedure only because the product includes a high-risk AI system as a safety component, if the product legislation does not require it.

Annex III systems. For biometrics (point 1), Article 43(1) allows internal control (Annex VI) only where harmonised standards or common specifications are applied; otherwise a notified body assesses the system under Annex VII. For points 2 to 8, Article 43(2) provides for internal control under Annex VI.

One declaration, one CE marking. Article 47 requires an EU declaration of conformity for each high-risk AI system, with the content set out in Annex V. Where the system is also subject to other Union harmonisation legislation requiring a declaration, a single EU declaration of conformity covers all of it (Article 47(3)). The CE marking follows Article 48; where a notified body is responsible for the Article 43 procedure, its identification number follows the marking (Article 48(4)).

Where the CRA fits (CRA Article 12, AI Act Article 42(3))

AI-enabled hardware with a direct or indirect data connection is also a product with digital elements under the CRA, unless an Article 2 exclusion applies. The CRA product requirements apply from 11 December 2027; Article 14 reporting has applied since 11 September 2026.

Article 12 of Regulation (EU) 2024/2847 links the two acts:

  • Article 12(1). A product with digital elements that is a high-risk AI system is deemed to comply with the cybersecurity requirements in Article 15 of the AI Act where it fulfils the essential cybersecurity requirements in Annex I Part I of the CRA, the manufacturer's processes fulfil Annex I Part II, and the EU declaration of conformity issued under the CRA demonstrates the level of cybersecurity protection that Article 15 requires. The accuracy and robustness requirements of Article 15 are outside this presumption. Regulation (EU) 2026/1744 mirrored the rule in the AI Act as Article 42(3).
  • Article 12(2). For those products and requirements, the conformity assessment procedure is the one in Article 43 of the AI Act.
  • Article 12(3). As an exception, important products (CRA Annex III) and critical products (CRA Annex IV) that would be assessed under AI Act internal control (Annex VI) follow the CRA's own conformity assessment procedures for the essential cybersecurity requirements.

The dates do not line up. The CRA product requirements apply from 11 December 2027, the AI Act high-risk rules for AI in products from 2 August 2028. Between those dates a connected product has to meet the CRA on its own terms.

Primary references

FAQ

Frequently asked questions

  • Does adding AI to a hardware product make it a high-risk AI system?

    Not by itself. Under Article 6(1) of Regulation (EU) 2024/1689 two conditions must both be met: the AI system is a safety component of a product (or is itself a product) covered by the legislation listed in Annex I, and that product must undergo third-party conformity assessment under that legislation. Article 6(1a), added by Regulation (EU) 2026/1744, states that AI used solely for non-safety aspects of user assistance, performance optimisation, service efficiency, automation, convenience or quality control is not a safety component. AI whose failure would endanger health and safety is one (Article 6(1b)).

  • When do the AI Act high-risk rules start to apply to AI embedded in products?

    From 2 August 2028 for AI systems that are high-risk under Article 6(1) and Annex I, and from 2 December 2027 for the Annex III use cases under Article 6(2). Regulation (EU) 2026/1744, published in the Official Journal on 24 July 2026 and in force since 27 July 2026, set these dates in Article 113 of Regulation (EU) 2024/1689. Chapters I and II, including the prohibited practices, have applied since 2 February 2025; the two prohibitions added in 2026 apply from 2 December 2026.

  • Does a notified body under the Radio Equipment Directive make my embedded AI high-risk?

    Not on that ground alone. Directive 2014/53/EU is listed in Annex I, Section A, of the AI Act, but Article 6(1c), added by Regulation (EU) 2026/1744, says a product that needs third-party conformity assessment solely because of risks other than health and safety — in particular radio-spectrum or electromagnetic-interference risks that do not affect health and safety — does not fulfil the third-party-assessment condition in Article 6(1)(b).

  • How do the AI Act and the Cyber Resilience Act interact on cybersecurity?

    Article 12(1) of Regulation (EU) 2024/2847 (CRA) deems a product with digital elements that is a high-risk AI system to comply with the cybersecurity requirements of Article 15 of the AI Act where it fulfils CRA Annex I Part I, the manufacturer's processes fulfil Annex I Part II, and the EU declaration of conformity issued under the CRA demonstrates the level of protection Article 15 requires. Regulation (EU) 2026/1744 mirrored this in Article 42(3) of the AI Act. Accuracy and robustness under Article 15 are not covered by this presumption.

Continue reading

Related guides

Put this into practice

Free tools & references

New to the terminology? Browse the compliance glossary — plain-English, citation-backed definitions of every term above.