EU AI Act
Also known as: European AI Act · EU artificial intelligence regulation
The EU AI Act sorts artificial intelligence systems by level of risk and imposes escalating obligations depending on the category a system falls into.
Its logic is a pyramid: some practices are prohibited, one category of systems is designated high-risk and carries most of the obligations, and other systems owe only transparency — telling a person they are dealing with a machine, or that content was generated.
It targets deployment as much as design. A company that trains no model but deploys one in a hiring or credit process carries its own obligations, distinct from the model provider's.
Like the GDPR, its reach is extraterritorial: where you are established matters less than whether the system's output is used in the Union.
What it means for a small business
The useful question for a small business is not "do I do AI" but "what is the system I deployed used for". An agent that drafts emails and an agent that screens job applications do not land in the same category, with the same technology underneath.
So what actually applies to you?
A definition tells you what a term means, not what your organization must do. The assessment answers the second question — free, no credit card.
Updated September 1, 2026 · Educational definition; not legal advice.