What Is High-Risk AI?
The EU AI Act, Regulation (EU) 2024/1689, follows a risk-based approach to artificial intelligence regulation.
Under Article 6, certain AI systems are classified as high-risk because of their intended purpose and potential impact on health, safety or fundamental rights.
Importantly, high-risk does not necessarily mean that a system is dangerous or already causing harm. It is a legal classification that triggers additional regulatory requirements.
The distinction matters because not every powerful or sophisticated AI system automatically qualifies as high-risk.
When Is an AI System Classified as High-Risk?
Article 6 establishes two principal routes to classification.
The first concerns AI systems used as safety components in certain regulated products, or constituting such products themselves. Under Article 6(1), the classification applies where the relevant EU product legislation listed in Annex I also requires third-party conformity assessment.
The second route concerns specific applications listed in Annex III.
These include certain AI systems used in recruitment, education, creditworthiness assessments, critical infrastructure, law enforcement and the administration of justice.
For example, software intended to evaluate job applicants may qualify as high-risk because its recommendations can significantly affect employment opportunities.
However, being used in a sensitive sector does not automatically make every AI application high-risk.
Article 6(3) provides exceptions for certain Annex III systems that do not pose a significant risk of harm, including systems performing narrow procedural or preparatory tasks under specified conditions.
These exceptions do not apply where the system performs profiling of natural persons. Providers relying on an exception must document their assessment and fulfil the applicable registration requirements.
What Requirements Apply to High-Risk AI?
High-risk classification brings substantial legal obligations.
Providers must ensure compliance with requirements concerning risk management, data governance, technical documentation, record-keeping, transparency, human oversight, accuracy, robustness and cybersecurity.
They must also complete the applicable conformity assessment before placing the system on the market or putting it into service.
Deployers have separate responsibilities, including using systems according to instructions and ensuring appropriate human oversight.
Following amendments introduced in 2026, the principal high-risk provisions will apply from 2 December 2027 for Annex III systems and 2 August 2028 for Annex I systems, subject to relevant transitional arrangements.
Final Thoughts
The high-risk classification demonstrates that the EU AI Act is concerned not simply with what artificial intelligence can do, but with where and how it is used.
A model that assists with routine administrative work may pose different legal risks when deployed to evaluate job applicants or influence judicial decisions.
Ultimately, responsible AI regulation depends on recognising that the consequences of a technology often matter more than its technical sophistication.
Sources & References
- Article AI Act: https://ai-legalinsight.com/ai-regulation/the-eu-ai-act-explained/
- Article 6: Classification for high-risk AI systems: https://ai-act-service-desk.ec.europa.eu/en/ai-act/article-6?