EU AI Act AI literacy training explained

Article 4 of the EU AI Act asks companies that provide or use AI systems to take measures to ensure, to their best extent, a sufficient level of AI literacy among their staff and others who use AI on their behalf. In practice that means training people on the tools they actually use, matched to their role, and keeping a record that you did it.

This guide is for HR, L&D, compliance and operations leads in the UK and the EU. It is general information, not legal advice. Confirm how the Act applies to your organisation with a lawyer.

Key facts about training your team with us

Live cohorts on Zoom that start every Monday, with seats for any number of team members
Invoiced to your company
Safe use of company data (GDPR) built in
An attendance register and certificates for your skills records
No coding needed

What does Article 4 of the EU AI Act say?

Article 4 is the AI literacy obligation. It applies to providers, who build or place AI systems on the market, and to deployers, who use AI systems in their work. It has applied since 2 February 2025, one of the first parts of the Act to take effect.

The Act does not prescribe a course, an exam or a number of hours. It asks for a sufficient level of AI literacy, taking into account people's technical knowledge, experience, education and training, and the context the AI systems are used in. You can read the text of the Act on EUR-Lex, the EU's official legal database.

Does it apply to companies that only use ChatGPT or Copilot?

Generally, yes. A company whose staff use general-purpose AI assistants for work is using AI systems, which makes it a deployer for the purposes of Article 4. The obligation is about the people using AI on your behalf, not only about high-risk systems.

UK companies are not in the EU, but the Act can still reach them, for example when they place AI systems on the EU market or when output of their AI systems is used in the EU. Many UK groups with EU staff or EU clients choose to meet the same standard across the business. Your lawyer can tell you where you stand.

What does AI literacy mean for an office worker?

For most staff, AI literacy is practical. A person who uses AI tools at work should be able to:

Explain in plain words what the tool does and where it tends to go wrong.
Write a clear request with context, a format and an example.
Check the output for invented facts, missing points and bias before using it.
Keep personal data and confidential information out of tools that are not approved for it, in line with the GDPR and your own policies.
Know who to ask, and what to do if something goes wrong.

Managers and people who choose or configure tools need more: how to judge a use case, where human review must stay in place, and how to document decisions.

Training a team? Your team members can join our live cohorts, which start every Monday, invoiced to your company, with safe use of company data (GDPR) built in and an attendance register and certificates as evidence of AI literacy training. Send a team enquiry.

What should AI literacy training include?

Good training is matched to the role and uses real work. Look for:

The tools your staff use. ChatGPT, Claude, Gemini or Copilot, compared on real tasks rather than described in slides.
Prompting with context. How the same request gives a weak or a strong answer depending on what you tell the tool.
Checking and accountability. Who signs off AI-assisted work, and how to check it.
Data protection. What may and may not go into each tool under your policy and the GDPR.
Limits and risks. Invented facts, outdated information, bias and over-reliance.
Practice. Each person completing a real task from their own job.

Our guide to AI training for employees goes into each of these in more depth.

What records should we keep as evidence?

The Act does not set a format, but it makes sense to be able to show what you did. Keep:

An attendance register with names, roles and dates.
A certificate or record of completion for each person.
A description of what the training covered.
Your AI use policy and the date each person received it.
A note of how you matched training to roles, for example extra sessions for managers.

Review the records once a year, and again when you introduce a new tool.

How often should staff be trained?

AI tools change every few months, so one session is a start, not an end. A practical rhythm is a solid first course for everyone who uses AI, a short refresher when a tool changes or a new one is approved, and onboarding for new joiners before they get access.

How do live cohorts work for UK and EU teams?

Our cohorts run live on Zoom, so time zones matter more than location. Team members join the same live sessions, work on their own tasks, and receive an attendance register and certificates you can keep as evidence. To see the style of teaching first, anyone can join the free weekend workshop, 10:00-13:00 UK or 11:00-14:00 CET.

If you are writing the policy that sits alongside the training, our guide to writing an AI policy is a useful starting point.

Is this legal advice?

No. This page is general information about the AI literacy obligation in the EU AI Act. How it applies depends on your organisation, your tools and where you operate. Confirm your obligations with a lawyer.

Frequently asked questions

What is the AI literacy requirement in the EU AI Act?

Article 4 asks providers and deployers of AI systems to take measures to ensure, to their best extent, a sufficient level of AI literacy among staff and others who use AI on their behalf.

Does Article 4 apply if we only use ChatGPT?

Generally yes. Using AI assistants for work makes a company a deployer, so the AI literacy obligation applies. Confirm your position with a lawyer.

What evidence should we keep?

Keep an attendance register, certificates, a description of the training, your AI use policy and a note of how training was matched to roles.