AI training at work: what the AI Act requires

What Article 4 says

Regulation (EU) 2024/1689, known as the AI Act, sets the rules for artificial intelligence in the European Union. Its Article 4 is called “AI literacy”, meaning the knowledge and skills needed to use AI with good judgement. It has applied since 2 February 2025.

It covers providers, who develop the tools, and deployers. A deployer is an organisation that uses an AI system under its own authority, except for personal, non-professional use. If an employee writes a quote or summarises a meeting with an AI assistant, your business is a deployer.

In the version in force since 27 July 2026, the article asks you to take “measures to support the development of AI literacy” of your staff. In plain terms: help everyone understand AI and use it well. The article also covers people who use AI on your behalf, such as a contractor. These measures take into account each person’s knowledge, experience and training. They also take into account the context of use and the people affected by the results. The text states that you do not have to guarantee a specific level of AI literacy for each person.

What the Digital Omnibus changed

Regulation (EU) 2026/1744, known as the “Digital Omnibus on AI”, has been in force since 27 July 2026. It reworded Article 4. Before, the text asked businesses to ensure, to their best extent, a sufficient level of AI literacy. That level requirement has gone, but the duty to act remains: you must still take measures. The European Commission and EU countries must also support these efforts, especially for SMEs. In particular, the Commission must publish practical examples.

The omnibus postponed the rules for “high-risk” systems to 2 December 2027, and to 2 August 2028 for some of them. This delay does not affect Article 4. Since 2 August 2026, national authorities have been able to check that it is applied.

What is not required

According to the European Commission’s questions and answers on AI literacy, updated on 27 July 2026:

• no certification is required;
• you do not have to measure each employee’s level;
• you do not have to appoint an AI officer.

However, the same document warns that, in many cases, simply relying on the tools’ instructions for use, or asking staff to read them, may not be enough. Plan concrete actions.

Four steps to get started

1. List how AI is used. Note the tools used, by whom and for which tasks: customer follow-ups, meeting notes, product descriptions, Excel spreadsheets.
2. Give everyone the same foundations. An AI assistant predicts a likely sequence of words: it does not “know”. It can produce answers that are wrong but convincing. Everyone also needs to know which data must never be entered. The Learn path covers these basics in 5-minute lessons, from level A1 to C2. The French government is also planning a free training platform, ExploreIA, from January 2027.
3. Adapt to each role. A salesperson who writes emails does not face the same risks as someone who sorts job applications. Plan specific guidance for sensitive uses.
4. Write simple rules. A one-page usage policy sets out the approved tools and the data that is off limits.

How to document it

The Commission’s questions and answers say that no certificate is needed: you can keep an internal record of training sessions and other support actions. Keeping track of the content, the date and the participants is good practice. A simple shared spreadsheet will do.

Also keep the materials used and the proof of completion, for example verified levels. Proof of completion is not a certification, but it shows that an action took place. To review your situation, use the self-assessment.