Reference · defined terms
EU AI Act glossary
Not every defined term, only the ones that change an outcome. Each entry names the article it comes from, because in this Regulation a definition is frequently the whole argument.
Art. 3 definitionsReg. (EU) 2024/1689 as amended
| Term | Source | Definition |
|---|---|---|
| AI system | Art. 3(1) | A machine-based system designed to operate with varying levels of autonomy, that may exhibit adaptiveness after deployment, and that infers from the input it receives how to generate outputs such as predictions, content, recommendations or decisions. Inference is the test. More → |
| Provider | Art. 3(3) | A person that develops an AI system or has one developed and places it on the market or puts it into service under its own name or trade mark, whether for payment or free of charge. More → |
| Deployer | Art. 3(4) | A person using an AI system under its own authority, except where use is a personal non-professional activity. More → |
| Intended purpose | Art. 3(12) | The use for which a system is intended by the provider, including context and conditions of use, as specified in the instructions for use, promotional or sales materials, and the technical documentation. More → |
| Safety component | Art. 3 | A component fulfilling a safety function for a product or system, or whose failure endangers health, safety or property. The gate for Annex III point 2 and the Annex I route. More → |
| Substantial modification | Art. 43(4) | A change triggering fresh conformity assessment. Changes pre-determined at initial assessment and described in the technical documentation are not substantial. More → |
| Serious incident | Art. 3(49) | An incident or malfunction leading directly or indirectly to death or serious harm to health, serious and irreversible disruption of critical infrastructure, infringement of Union law protecting fundamental rights, or serious harm to property or the environment. More → |
| Profiling | GDPR Art. 4(4) | Automated processing to evaluate personal aspects of a natural person. Removes the Article 6(3) derogation outright. More → |
| Presumption of conformity | Art. 40 | A rebuttable presumption that a system conforming to a harmonised standard cited in the Official Journal meets the requirements that standard covers. No AI Act harmonised standard has been cited yet. More → |
| Harmonised standard | Art. 40 | A European standard whose reference has been published in the Official Journal. ISO/IEC 42001 is not one. More → |
| Notified body | Ch. III Sec. 4 | A conformity assessment body designated by a notifying authority. Required for Annex III point 1 biometrics where harmonised standards have not been applied. More → |
| GPAI model | Ch. V | A general-purpose AI model displaying significant generality, capable of competently performing a wide range of distinct tasks. Governed by Articles 51–56. More → |
| Systemic risk | Art. 51 | A classification for GPAI models with high impact capabilities, presumed on cumulative training compute and amendable by delegated act. More → |
| FRIA | Art. 27 | Fundamental rights impact assessment. A deployer obligation for public bodies, private entities providing public services, and deployers of Annex III 5(b) and 5(c) systems. More → |
| Market surveillance authority | Art. 70, 74 | The national body enforcing most of the Regulation. Powers took effect 2 August 2026. More → |
| AI Office | Art. 64 | The Commission body supervising GPAI models, with exclusive competence after Reg. 2026/1744 over systems built on a provider’s own GPAI model and over DSA-designated VLOPs and VLOSEs. More → |
| Digital Omnibus on AI | Reg. (EU) 2026/1744 | The first amendment to the AI Act. OJ 24 July 2026, in force 27 July 2026. More → |
Three definitions do most of the work
AI system decides whether the Regulation applies at all. Intended purpose decides classification, and it expressly includes your marketing copy. Profiling decides whether the Article 6(3) derogation is available, and for most systems that evaluate people, it is not.