Skip to main content
Praxikon
Agent check
Article 3 of 1133%
Nederlands

Article 3: Definitions

According to Article 3, an AI system is a machine-based system designed to operate with varying levels of autonomy, that may exhibit adaptiveness, and infers from input how to generate outputs such as predictions, recommendations or decisions.

, AI and privacy legal adviserobligations checked against the source on
Check your own case for free

Praxikon tracks Article 3 (Definitions) under the EU AI Act, citing the source for every statement.

EU Official:
Title I: General Provisions

Official text

||

Source: EUR-Lex, Regulation (EU) 2024/1689. Text reproduced verbatim.

Download AI Act (PDF) →

Decision tree

Is your application an AI system?

Assess the system as it actually works, not as it is called in marketing. A product can contain AI components next to ordinary software.

Question 1 · Art. 3(1)

Does the system itself infer, from the input it receives, how to generate outputs, for example through machine learning or logic- and knowledge-based techniques?

Software that only executes rules defined entirely by people does not infer anything. The guidelines also give basic data processing, classical heuristics and systems that only improve mathematical optimisation as examples outside the definition.

Full diagram
  1. 1

    Does the system itself infer, from the input it receives, how to generate outputs, for example through machine learning or logic- and knowledge-based techniques?

    Art. 3(1)

    • Yes, it infers → Question 2
    • No, it only executes fixed rules → Outcome B
  2. 2

    Is the system designed to operate with some level of autonomy, meaning some independence from human involvement in each step?

    Art. 3(1)

    • Yes → Question 3
    • No, a human steers every step → Outcome B
  3. 3

    Does the system, for explicit or implicit objectives, generate outputs such as predictions, content, recommendations or decisions that can influence physical or virtual environments?

    Art. 3(1)

    • Yes → Question 4
    • No → Outcome B
  4. 4

    Do you place the system on the market or put it into service in the EU, use it from an establishment in the EU, or is its output used in the EU?

    Art. 2(1)

    • Yes → Question 5
    • No, no link with the EU at all → Outcome C
  5. 5

    Is the system used exclusively for military, defence or national security purposes?

    Art. 2(3)

    • Yes, exclusively → Outcome D
    • No → Question 6
  6. 6

    Is the system specifically developed and put into service for the sole purpose of scientific research and development?

    Art. 2(6)

    • Yes, solely for that → Outcome E
    • No → Question 7
  7. 7

    Is the system still in research, testing or development and not yet placed on the market or put into service, and are you not testing it in real world conditions?

    Art. 2(8)

    • Yes → Outcome F
    • No, it is in use or tested in real world conditions → Question 8
  8. 8

    Do you use the system as a natural person in the course of a purely personal, non-professional activity?

    Art. 2(10)

    • Yes → Outcome G
    • No, we use or supply it professionally → Question 9
  9. 9

    Is the system released under a free and open-source licence?

    Art. 2(12)

    • Yes → Outcome H
    • No → Outcome A

Outcome A · Applies

It is an AI system within the scope of the AI Act

The regulation applies. Which obligations apply depends on your role (provider or deployer) and on the risk: first the prohibited practices, then the high-risk classification, then the transparency obligations. Article 4 on AI literacy applies to every provider and deployer.

Art. 3(1) · Art. 2(1) · art. 4 · art. 5 · art. 6 · art. 50

Outcome B · Does not apply

Probably not an AI system

Without the capacity to infer, some autonomy or outputs that can influence environments, the application falls outside the definition. Record the reasoning, because the line is often thin in practice. If the application processes personal data, the GDPR still applies.

Art. 3(1) · overweging 12

Outcome C · Does not apply

Outside the territorial scope

Without placing on the market or putting into service in the EU, without a deployer in the EU and without use of the output in the EU, the regulation does not apply. If the output is later used in the EU, that changes.

Art. 2(1)

Outcome D · Does not apply

Excluded: military, defence or national security

The regulation does not apply to AI systems placed on the market, put into service or used exclusively for military, defence or national security purposes, regardless of who does so. If the system is also used for other purposes, the regulation applies to that use.

Art. 2(3)

Outcome E · Does not apply

Excluded: solely scientific research

Systems and models specifically developed and put into service for the sole purpose of scientific research and development, including their output, fall outside the regulation. If the system later goes to market or into use for another purpose, the regulation does apply.

Art. 2(6)

Outcome F · Partly applies

Not yet: the development phase falls outside the regulation

Research, testing and development before placing on the market or putting into service fall outside the regulation, except testing in real world conditions. Other Union law, such as the GDPR, still applies. As soon as you place the system on the market or put it into service, the obligations apply.

Art. 2(8)

Outcome G · Does not apply

No obligations for you as a private user

Deployer obligations do not apply to natural persons using an AI system in the course of a purely personal, non-professional activity. The provider of the system keeps its own obligations.

Art. 2(10) · Art. 3(4)

Outcome H · Partly applies

Open source: only for high risk, prohibited practices or transparency duties

The regulation does not apply to AI systems released under free and open-source licences, unless they are placed on the market or put into service as high-risk AI systems, or fall under Article 5 or Article 50. The exclusion attaches to the system, not to your role: it therefore also covers you if you only use the system. If you deploy it for a high-risk purpose, for example one listed in Annex III, it is put into service as a high-risk system and the exclusion falls away; then check under Article 25 who is the provider. If the system is provided against a price or otherwise monetised, for example through paid support, the exclusion does not apply according to recital 103. General-purpose AI models have their own regime (Article 53(2)).

Art. 2(12) · art. 5 · art. 6 · art. 50 · art. 25 · overweging 103

A decision tree simplifies. If you are unsure at a question, read the provision or have it assessed. Checked against the legal text on 1 October 2026.

Official guidance on this article

7

Related enforcement

No enforcement actions for this article yet. Follow developments via the Enforcement Tracker.

Related blog posts

Related articles

Frequently asked questions

How does Article 3 AI Act define an AI system?+
According to Article 3, an AI system is a machine-based system designed to operate with varying levels of autonomy, that may exhibit adaptiveness, and infers from input how to generate outputs such as predictions, recommendations or decisions.
What is a provider under the AI Act?+
Article 3(3) defines a provider as a person or organisation that develops an AI system or general-purpose AI model and places it on the market or puts it into service under its own name or trademark.
What is a deployer under Article 3 of the AI Act?+
Article 3(4) defines a deployer as a person or organisation using an AI system under its authority, except where the AI system is used in the course of a personal non-professional activity.

Connections

What connects to Article 3 AI Act

Themes where this returns

Where this meets the other law

The counterpart in the other law

GDPR interpretation that also applies here6 of 15