Legal news
The inference capacity of AI systems : key element of their definition
On February 6, 2025, the European Commission published guidelines on the notion of "AI system" [1] ("Guidelines") as defined by article 3(1) of Regulation (EU) 2024/1689 on Artificial Intelligence [2] ("AI Act") and clarified by its Recital 12.
As a reminder, the AI Act defines an AI system 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".
The publication of these Guidelines results from the application of article 96 of the AI Act, which states that "The Commission shall develop guidelines on the practical implementation of this Regulation, and in particular on : [...] f) the application of the definition of an AI system as set out in Article 3, point (1)".
As the Commission points out, the AI Act applies only to AI systems that meet the definition set out in its article 3. Consequently, and given that the definition of AI system and of certain provisions of Chapters I "General Provisions", and II "Prohibited AI Practices", of the AI Act came into force on February 2, 2025, it was essential that the Commission clarifies this crucial concept for the various players concerned by the Act.
In line with the definition provided by the AI Act, the Guidelines identify the following seven elements characterizing an AI system :
1. a machine-based system
2. that is designed to operate with varying levels of autonomy
3. that may exhibit adaptiveness after deployment
4. and that, for explicit or implicit objectives
5. infers, from the input it receives, how to generate outputs
6. such as predictions, content, recommendations, or decisions
7. that can influence physical or virtual environments
With regard to the inference capacity of AI systems, a key element in the definition of these systems, the Guidelines provide examples and details of AI techniques offering such inference capacity, and give examples of systems that do not fall within the scope of the definition of AI system (e.g. basic data processing following predefined and explicit instructions).
Qualification on a case-by-case basis
As reminded by the Commission, the qualification of a machine-based system with regard to the definition of AI system must be carried out on a case-by-case basis, taking into account the architecture and functionality of each system and the seven elements above mentioned. Consequently, there can be no automatic determination or exhaustive list of systems constituting, or not, AI systems.
The Guidelines point out that they are not binding, but that only a decision by the Court of Justice of the European Union can give an authoritative interpretation of the AI Act.
Read more on the subject of AI :
[1] Commission Guidelines on the definition of an artificial intelligence system established by Regulation (EU) 2024/1689 (AI Act)