On this page
  1. Who and since when
  2. Example
  3. What it means for you
  4. Read next

Last verified October 31, 2026, against the AI Act as amended by the Digital Omnibus on AI (Regulation (EU) 2026/1744).

Where it comes from. The definition is in Art. 3(56) of the EU AI Act. The duty is in Art. 4(1): providers and deployers “shall take measures to support the development of AI literacy of their staff and other persons dealing with the operation and use of AI systems on their behalf”.

Who and since when

Every provider and deployer in scope, whatever the system’s risk tier, including tools often called minimal risk. Art. 4 has applied since February 2, 2025 (Art. 113). The Digital Omnibus amended it with effect from July 27, 2026: the text now says the duty “does not require providers or deployers to guarantee any specific level of AI literacy of any individual”.

Example

A US marketing agency uses a generative AI tool to draft ad copy for its clients in the EU, so the tool’s output is used in the EU and the agency is a deployer (Art. 2(1)(c)). It gives its staff, and the freelancers who use the tool on its behalf, a short session on what the tool does, where it gets things wrong, and how to check output before it goes out.

What it means for you

  • Fit the measures to the people and the context: their technical knowledge, experience, education and training, the context of use, and the people the systems are used on (Art. 4(1)).
  • The duty covers your staff and others who operate or use AI systems on your behalf, such as contractors. The text does not name your customers or end users.
  • The Act sets no format, certificate or course. Keeping a short record of the measures you took can help you show them.
  • Art. 4(2) asks the Commission to publish practical examples of how to meet the duty.

This page is general information, not legal advice.

All glossary entries

Questions about your AI systems?

Talk to us about what you build or use, or see how an assessment works.

How it works