Skip to content

Guide

AI literacy training: what Article 4 of the EU AI Act requires

In short

GustinAI's AI training for companies helps employers respond to Article 4 of the EU AI Act, which requires developing staff AI literacy. This guide explains what the text says, what it does not require, and how practical training answers it in practice. It is not legal advice.

Updated on

For more on putting this into practice, see my AI training for companies.

What Article 4 actually says

The EU's AI regulation, Regulation (EU) 2024/1689, is coming into application in stages. Its Article 4 requires providers and deployers of AI systems to take measures to ensure, to their best extent, a sufficient level of AI literacy of their staff and other persons dealing with the operation and use of AI systems on their behalf. This obligation has been applicable since 2 February 2025.

The text does not set a required format, a minimum duration, or a mandatory certification. It requires an outcome: that people who use AI at work understand its possibilities and its limits. This guide does not replace legal advice: to assess your specific situation, consult a lawyer specialising in digital law.

What Article 4 does not require

Contrary to a common assumption, Article 4 does not set a specific fine, does not require a set number of training hours, and does not list authorised tools. It also does not apply in the same way depending on whether a company is a provider or simply a deployer of an AI system. Implementation details and enforcement are matters for the relevant national authorities, which have not yet published detailed guidance on this point.

QuestionWhat Article 4 saysWhat it does not say
Who is coveredProviders and deployers of AI systemsNo stated headcount threshold
Since whenApplicable since 2 February 2025No published enforcement timeline
What level of trainingA "sufficient" level, to best extent possibleNo set number of hours
PenaltiesNot specified in the articleNo published figure to date

Why real training still matters, beyond compliance

Beyond the text, training has a measurable effect. In BCG's AI at Work 2025 survey of more than 10,600 employees in 11 countries, only 36% of employees feel properly trained in AI, and 79% of those who received more than five hours of training use AI regularly, compared with 67% of the others. A well-designed training therefore serves two purposes at once: demonstrating an AI literacy effort, and producing teams that genuinely use it.

How to build a credible response to Article 4

  1. Actually train your teams. A short, practical session is worth more than an internal memo with no follow-up.

  2. Document the process. Dates, content and participants of training sessions.

  3. Set written usage rules. An AI usage policy formalises what teams can and cannot do.

  4. Target the most exposed roles first. Teams who use AI daily, such as customer service or marketing, need it first.

Format

This training is delivered in-house, as a half day or a full 7-hour day, online or on-site on request. See the pricing page for details. To choose the right format for your team, also see the guide How to choose AI training.

Back to the full training

See AI training for companies

Frequently asked questions

Is this guide legal advice?

No. It presents the text of Article 4 for information purposes only. To assess your situation, consult a specialist lawyer.

Is my company covered if it only uses ChatGPT?

Article 4 covers providers and deployers of AI systems broadly. A lawyer can clarify how this text applies to your case.

Does online training satisfy Article 4?

The text does not set a required format. What matters is that the people concerned gain real AI literacy, whatever format is chosen.

What happens if a company does not train its staff?

Article 4 does not specify a particular penalty. Enforcement is a matter for the relevant national authorities.

Would you like to know how AI can help your company?

Book a free 30-minute diagnosis: I'll work with you to identify where artificial intelligence can save your team time.