Draft Guidelines on the Classification of AI Systems: How to Properly Determine the Type of Syste

27.05.2026

On May 19, 2026, the Draft Commission guidelines on the classification of high-risk AI systems were published. The guidelines are intended to assist providers and developers of AI systems, as well as competent market surveillance authorities, in assessing whether an AI system should be classified as a high-risk system. This should contribute to the uniform application and effective enforcement of Article 6 of the EU Artificial Intelligence Act (EU AI Act)

Key Aspects

 

First, the guidelines apply exclusively to AI systems within the meaning of the EU AI Act:Not every software product can automatically be classified as an AI system. Compliance with specific criteria must be verified.An AI system is defined as a machine-based system that is designed to operate with varying levels of autonomy and that may exhibit adaptiveness after deployment, and that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments.

Accordingly, the main criteria are:

 

1. Machine-based nature;
2. Capability to adapt after deployment;
3. Ability to generate outputs/data (for both explicit and implicit objectives);
4. Influence on physical or virtual environments.

 

Second, it is necessary to determine the intended purpose of the AI system:

This refers to the use for which an AI system is intended by the provider, including the specific context and conditions of use, as specified in the information supplied by the provider in the instructions for use, promotional or sales materials and statements, as well as in the technical documentation.

It should be remembered that a formal disclaimer excluding high-risk use of an AI system (for example, in the Terms of Service) will not be sufficient if the overall product presentation — including marketing positioning and examples of use in instructions — in fact allows for high-risk use cases.

Therefore, the product must be objectively assessed, and the entire set of materials should be reviewed in order to avoid liability risks arising from incorrect classification of the AI system.

 

What Should Be Considered Under Annex I (Article 6(1))?

If the system is intended to be used as a safety component of a product, or if the AI system itself is a product covered by EU harmonization legislation listed in Annex I, and the product whose safety component includes the AI system - or the AI system itself - must undergo third-party conformity assessment, the system will be classified as a high-risk system under Article 6(1).

What Is Defined in Annex III (Article 6(2))?

Annex III contains a list of areas and specific categories that are presumed to be high-risk by default (for example: biometrics, critical infrastructure, education and vocational training, employment, workers’ management and access to self-employment, administration of justice, and other sectors).

An AI system provider who believes that a specific system listed in Annex III should not be considered high-risk on these grounds must document the justification and provide such documentation upon request by the regulator.
 

Recommendations

Before deployment, an audit of the AI system should be conducted to assess compliance with high-risk criteria.
Even if an AI system is not initially intended for high-risk use, the following should still be considered:

• For general-purpose AI systems, it is mandatory to take into account the actual scope of the declared intended purpose and reasonably foreseeable use scenarios;
• For multi-purpose systems and general-purpose AI systems, it is important that limitations on use are clearly and consistently documented. High-risk use cases must be expressly excluded; otherwise, such use may become part of the intended purpose.

Why Is This Important for Business?

 

• Ensuring compliance with legal requirements;
• Minimizing liability risks;
• Facilitating timely market entry for AI products.

Incorrect classification of an AI system or inaccurate product descriptions in promotional materials may result in fines and the blocking of the product on the EU market.