Skip to main content

Annex I High-Risk AI Systems: The Product-Safety Route

Which AI systems in regulated products are high-risk under Article 6(1), which rules apply, and what changes for machinery.

Annex I lists EU product-safety laws. Providers of AI systems that are, or are safety components of, products under these laws must meet the high-risk requirements. The product must also need a third-party conformity assessment (by an outside body). The requirements apply from 2 August 2028.

This covers Annex I Section A. For Section B, including machinery, only Articles 6(1), 60a and 102 to 112 apply. Fines for breaking provider duties (Article 16) reach EUR 15 million or 3% of worldwide annual turnover, whichever is higher (Article 99(4)).

First check whether your product’s law is in Section A.

Annex I lists 20 laws in two sections, and the sections work differently

Section A has 11 laws (points 2 to 12) and Section B has 9 laws (points 13 to 21).

Section A: laws based on the New Legislative Framework

  • Toys (Directive 2009/48/EC)
  • Recreational craft and personal watercraft (Directive 2013/53/EU)
  • Lifts and safety components for lifts (Directive 2014/33/EU)
  • Equipment and protective systems for potentially explosive atmospheres (Directive 2014/34/EU)
  • Radio equipment (Directive 2014/53/EU)
  • Pressure equipment (Directive 2014/68/EU)
  • Cableway installations (Regulation (EU) 2016/424)
  • Personal protective equipment (Regulation (EU) 2016/425)
  • Appliances burning gaseous fuels (Regulation (EU) 2016/426)
  • Medical devices (Regulation (EU) 2017/745)
  • In vitro diagnostic medical devices (Regulation (EU) 2017/746)

Section B: other Union harmonisation legislation

  • Civil aviation security (Regulation (EC) No 300/2008)
  • Two- or three-wheel vehicles and quadricycles (Regulation (EU) No 168/2013)
  • Agricultural and forestry vehicles (Regulation (EU) No 167/2013)
  • Marine equipment (Directive 2014/90/EU)
  • Rail system interoperability (Directive (EU) 2016/797)
  • Motor vehicles and their trailers (Regulation (EU) 2018/858)
  • Motor vehicles, general safety type-approval (Regulation (EU) 2019/2144)
  • Civil aviation, for unmanned aircraft and their parts (Regulation (EU) 2018/1139)
  • Machinery (Regulation (EU) 2023/1230)

Regulation (EU) 2026/1744 removed the Machinery Directive 2006/42/EC from Section A and added the Machinery Regulation (EU) 2023/1230 to Section B.

Read the consolidated AI Act on EUR-Lex

An AI system is high-risk under Article 6(1) only if two conditions are both met

Both conditions must be met:

  1. The AI system is, or is intended to be used as a safety component of, a product covered by an Annex I law.
  2. That product must need a third-party conformity assessment before it is placed on the market or put into service under the Annex I law.

A safety component fulfils a safety function: its intended purpose is to prevent or mitigate risks to the health and safety of people or property (Article 3(14)).

AI used only for non-safety aspects of user assistance, performance optimisation, service efficiency, automation, convenience or quality control is not a safety component. If its failure would endanger health and safety, it is one (Article 6(1a) and (1b)).

A third-party assessment required only for risks other than health and safety does not meet condition 2. Examples are radio spectrum use and electromagnetic interference (Article 6(1c)).

Classifying a product as high-risk does not change which conformity assessment procedure its Section A law allows (Article 43(3)).

Section A: the high-risk rules apply from 2 August 2028

From 2 August 2028, Articles 6 to 27 (except Article 6(5)) apply to AI systems that are high-risk under Article 6(1) and a Section A law (Article 113). They set the requirements for high-risk AI systems (Articles 8 to 15) and the duties of providers, deployers and other parties (Articles 16 to 27). Systems placed on the market or put into service before that date are covered only after significant design changes; in any case, providers and deployers of high-risk systems intended to be used by public authorities must take the necessary steps to comply by 2 August 2030 (Article 111(2)).

Each requirement and the main provider duties have their own guide:

The conformity assessment follows the procedure of the Section A law, and the Chapter III requirements become part of it (Article 43(3)).

Providers may build the testing, reporting and documentation for these requirements into the documents and procedures they already keep under the Section A law (Article 8(2)).

For these systems, Articles 9 to 15 and 17 to 25 may be limited where and to the extent that a Section A law gives an equivalent or higher level of protection. The limit must not reduce the overall level of protection under the AI Act (Article 2(13)). By 2 August 2027, the Commission must adopt delegated acts that specify the systems concerned, the requirements that may be limited, the conditions and the scope.

Section B: only Articles 6(1), 60a and 102 to 112 apply

For an AI system that is high-risk under Article 6(1) and relates to a Section B product, only Article 6(1), Article 60a and Articles 102 to 112 apply (Article 2(2)).

Articles 57, 58 and 59 (AI regulatory sandboxes) apply only so far as the sector law has integrated the high-risk requirements.

Article 3(1) of Regulation (EU) 2026/1744 adds paragraphs to Article 8 of the Machinery Regulation (EU) 2023/1230. The Commission must adopt delegated acts that add health and safety requirements for AI systems. They cover AI systems that are high-risk under Article 6(1) because they are a safety component of machinery or are machinery themselves. Those delegated acts must apply by 2 August 2028.

The provider must act, and the provider can be the product manufacturer

The Act applies to providers that place AI systems on the Union market or put them into service, wherever the provider is established (Article 2(1)(a)).

For a high-risk AI system that is a safety component of a Section A product, the product manufacturer is the provider if the system is placed on the market with the product under the manufacturer’s name or trademark, or is put into service under that name or trademark after the product has been placed on the market (Article 25(3)).

The fine cap for provider duties depends on company size

Article 99(4) sets fines for breaking the provider obligations in Article 16. The cap depends on company size:

  • Article 99(4)
    Large company (not an SME or small mid-cap enterprise)

    Up to EUR 15 million or, for a company, 3% of total worldwide annual turnover of the preceding financial year, whichever is higher.

  • Article 99(6)
    SME, including start-ups

    Up to EUR 15 million or 3% of total worldwide annual turnover of the preceding financial year, whichever is lower.

  • Article 99(6a)
    Small mid-cap enterprise (SMC)

    Up to EUR 15 million or 3% of total worldwide annual turnover of the preceding financial year, whichever is lower.

Article 3 defines SMEs by Recommendation 2003/361/EC and small mid-cap enterprises by point (2) of the Annex to Recommendation (EU) 2025/1099.

The same cap covers authorised representatives, importers, distributors and deployers who break their own duties (Article 99(4), points (b) to (e)).

Giving notified bodies or national authorities incorrect, incomplete or misleading information on request is capped at EUR 7.5 million or, for a company, 1% of total worldwide annual turnover of the preceding financial year, whichever is higher (Article 99(5)). The SME and small mid-cap enterprise rule above applies here too (Article 99(6) and (6a)).

Article 99 is not on the list of articles that apply to Section B systems, so these fines do not apply to them (Article 2(2)).

Four checks show where your product stands

  1. Find the Annex I law that covers your product, and note whether it is in Section A or Section B.
  2. Check whether that law requires a third-party conformity assessment for your product.
  3. Check whether your AI system is the product or a safety component of it (Article 6(1) and Article 3(14)).
  4. For Section A, plan the Chapter III requirements (Sections 1 to 3) for 2 August 2028. For Section B, ask counsel which sector rules apply to your AI system.

Common questions about Annex I high-risk systems

Does the Annex I date apply to machinery?

Not directly. Machinery moved to Section B of Annex I, so only Article 6(1), Article 60a and Articles 102 to 112 apply to AI systems in machinery that are high-risk under Article 6(1) (Article 2(2)). The machinery regulation is to carry the AI requirements, through delegated acts that must apply by 2 August 2028 (Article 3(1) of Regulation (EU) 2026/1744).

Is every AI system in a product with an Annex I law high-risk?

No. Both conditions of Article 6(1) must be met. The AI system must be the product or a safety component of it, and the product must need a third-party conformity assessment. AI used only for convenience, optimisation or quality control is not a safety component unless its failure would endanger health and safety (Article 6(1a) and (1b)).

What if my AI system is also an Annex III use?

If the system is covered by a Section A law and also falls under Annex III, the provider follows the conformity assessment procedure of the Section A law (Article 43(3)).

Scanara supports the evidence work, and you decide the classification

Scanara scans code for standard ML framework patterns, structures each system’s classification and prepares evidence for the technical documentation. You decide whether your product falls under an Annex I law. A notified body, not Scanara, carries out any third-party conformity assessment.

See Scanara’s features

Read next: the Article 6 rules, Annex III and the provider duties