Knowledge base
Article 4 after the Digital Omnibus: what should an SME record?
Published on
Since 27 July 2026, AI literacy is an obligation of means. What that means in practice for a Belgian SME, and what to document.
What changed
Article 4 of the AI Act has applied since 2 February 2025. In the original text, providers and deployers of AI systems had to take measures to ensure, to their best extent, a sufficient level of AI literacy of their staff.
The Digital Omnibus, Regulation (EU) 2026/1744, entered into force on 27 July 2026 and rewrote that article. Organisations must now take measures to support the development of AI literacy, tailored to the knowledge, experience, education and context in which staff use AI. The text states that no specific level per person has to be guaranteed. The Commission and Member States are tasked with helping organisations, especially SMEs.
What did not change
The duty still exists. Anyone using AI systems, including everyday tools like ChatGPT, Microsoft 365 Copilot or Gemini, must take measures. SMEs are not exempt.
There is still no mandatory certificate, no training of a set length and no fixed format.
What to record
Because it is an obligation of means, the question after an inspection or incident becomes: which measures did you take, for whom and when? A simple register is enough. For each measure note the date, what you did, for which department or role, how many people, which AI tools, where the evidence is and the next step.
Typical measures are rules on which data may go into which tool, a short training, a measurement of what staff actually do, and a repeat a few months later. Our free record template has exactly those columns.
Why measuring is stronger than training alone
A list of completed courses shows that a measure existed. A measurement before and a second one after also show whether it fitted the people and their work, which is what the new text asks. They also help you aim training at the departments that really need it.
This month, in practice
List the AI tools used in your company. Agree which data may not go into them. Pick one department that uses AI a lot and measure where it goes well and where it doesn't. Put everything in your register. General information, not legal advice.