High-risk AI system
Also known as: high risk AI · Annex III system
A high-risk AI system is one whose use may harm health, safety or fundamental rights, and which therefore carries the bulk of the regulatory obligations.
The classification depends on use, not technology. The same language model is out of scope when summarising meeting notes and in scope when ranking job applications.
The domains classically covered are employment, education, access to credit and essential services, justice, migration, critical infrastructure, and certain biometric systems.
The attached obligations are heavy: risk management, data quality, technical documentation, logging, transparency to the user, human oversight, robustness and accuracy.
What it means for a small business
Before asking how to comply, establish whether you are in the category at all. Many companies worry about obligations that do not reach them while a genuinely in-scope use goes unnoticed.
Related terms
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.