Definition · Plain-language
High-risk AI system
A high-risk AI system, under the EU AI Act, is an AI system used in sensitive Annex III domains or acting as a safety component, subject to the Act’s strictest requirements.
The step most authors miss
Doing CRediT right? Don’t stop at the statement.
A CRediT statement credits you inside one paper. The recognition CRediT was built for happens when those roles are tied to you, persistently. Sign in with your ORCID — free — and claim your CRediT contributions on casrai.org, the home of the standard. They become a verified, portable part of your identity, not a line that disappears into one PDF.
Free: claim your contributions, then export a journal-ready CRediT statement, schema.org structured data, JATS XML, CSV or BibTeX — and preview your public profile. A membership publishes that profile publicly and verifies the journals you serve.
How a system becomes high-risk
The EU AI Act identifies high-risk AI systems through two main routes. The first covers AI systems that are themselves products, or safety components of products, already regulated under EU harmonised legislation requiring third-party conformity assessment — for instance certain machinery, medical devices or vehicles. The second covers AI systems used in the specific areas listed in Annex III. A system falling into these categories is treated as high-risk and must meet the Act’s detailed obligations, although a mechanism allows a system listed in Annex III to be considered not high-risk where it does not pose a significant risk.
The Annex III domains
Annex III sets out the use areas that make a system high-risk. They include certain biometrics; critical infrastructure such as the safety of road traffic or supply of utilities; education and vocational training, for example systems determining access or evaluating learning outcomes; employment and worker management, including recruitment and decisions on promotion or termination; access to essential private and public services, such as creditworthiness assessment and certain insurance; law enforcement; migration, asylum and border control; and the administration of justice and democratic processes. These are domains where AI could materially affect rights and life chances.
Requirements for high-risk systems
High-risk AI systems must satisfy a defined set of requirements before and after they reach the market. These include establishing a risk-management system, applying data-governance measures to training, validation and testing data, drawing up technical documentation, enabling record-keeping through logging, providing transparency and information to deployers, ensuring effective human oversight, and achieving appropriate accuracy, robustness and cybersecurity. Providers must carry out the relevant conformity assessment, draw up an EU declaration of conformity and affix CE marking. Most of these obligations apply from August 2026 under the Act’s phased timeline.
Key facts
At a glance
- Definition: AI in Annex III domains or a safety component, under the EU AI Act.
- Two routes: Regulated-product safety components, or Annex III use areas.
- Annex III: Employment, education, essential services, biometrics, law enforcement, more.
- Core duties: Risk management, data governance, human oversight, documentation.
- Market step: Conformity assessment + EU declaration + CE marking.
- Applies from: Largely August 2026 (some to 2027).
Common misconceptions
What people often get wrong
Often heard: Any powerful or advanced AI system is automatically high-risk.
Actually: High-risk status depends on use, not raw capability. A system is high-risk because it falls within Annex III domains or is a regulated-product safety component, not simply because it is technically advanced.
Often heard: High-risk AI systems are banned under the EU AI Act.
Actually: High-risk systems are permitted but heavily regulated. They must meet strict requirements and undergo conformity assessment. The banned category is the separate set of prohibited practices under Article 5.
Often heard: Every system listed in Annex III is always treated as high-risk.
Actually: The Act includes a mechanism by which a system in an Annex III area may be considered not high-risk where it does not pose a significant risk to health, safety or fundamental rights, subject to conditions.
Going deeper
