Glossary · EU AI Act: foundations
AI system (AI Act)
Also known as: AI system, artificial intelligence system
Under the EU AI Act, an AI system is a machine-based system designed to operate with varying levels of autonomy, which may exhibit adaptiveness after deployment and which, 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.
- AI regulation
- EU
In one sentence
Under the AI Act, an AI system is a machine-based system that infers from its input how to generate predictions, content, recommendations or decisions.
Example
A chatbot answering questions from a product manual is an AI system; a rule-based troubleshooting wizard with fixed decision trees written by people is not.
How it applies
- Scope test: The definition decides whether the AI Act applies at all. The key element is inference: the system derives from input how to generate outputs, rather than only executing rules defined by people. The Commission published guidelines on the definition in February 2025.
- Technical documentation: Describe in the documentation how the system infers — which model, which learning approach — because that is what makes it an AI system. Conventional software with purely human-defined rules is outside the definition.
- Origin: The wording is based on the OECD definition of an AI system, which is also used in many non-EU frameworks.
Compared with the USA, Canada and China
The US NIST AI RMF and Canada's former AIDA also used definitions close to the OECD wording. China's rules define their scope by technology instead — algorithmic recommendation, deep synthesis, generative AI services — so the question is not “is it AI?” but “which service is it?”.