What is a high-risk AI system?

The AI Act classifies systems by level of risk. Some uses are prohibited, many are minimal risk and remain free, and others carry only transparency obligations. High-risk systems require risk management, quality data, documentation, event logging, human oversight and demonstrated accuracy. The obligations are heavier for whoever provides the system, but anyone who uses it must also follow its instructions and ensure human control.

According to the Commission, after the changes made by the Digital Omnibus on AI, the rules for systems used in high-risk areas such as employment and education apply from 2 December 2027, and those for systems built into products from 2 August 2028 ¹. The Digital Omnibus is Regulation (EU) 2026/1744, in force since 27 July 2026 ².

An example in an SMB

A logistics company with many warehouse staff is considering software that ranks candidates and suggests who to call for an interview. Giovanna, the HR manager, finds out that it counts as a high-risk use. She asks the supplier for the compliance documentation and decides that every final choice stays with a trained person.

The most common mistakes

• Thinking that high risk only concerns large tech companies.
• Believing that a general-purpose chatbot used to write emails is automatically high-risk.
• Buying a system for selecting or evaluating workers without asking the supplier for the documentation.

Why it matters to you

If you use or are thinking of using AI for decisions about workers, students, credit or safety, you are in high-risk territory. Better to know before you sign a contract.

All the words of AI explained simply are in the glossary.