Bern
Lisbon
New York
What AI literacy training do staff need?
Staff need AI literacy training matched to how they use AI: everyone needs the basics, while people who own, build or decide on AI use cases need more. The EU AI Act describes AI literacy as the skills, knowledge and understanding that allow people to use AI systems in an informed way and to be aware of their opportunities, risks and possible harms (AI Act, Art. 3).
What the law requires
Legal requirement (EU AI Act, Art. 4). Since 2 February 2025, providers and deployers of AI systems must take measures to support the development of AI literacy (AI Act, Art. 4). These measures cover not only staff but also other persons dealing with the operation and use of AI systems on the company’s behalf, such as contractors and freelancers. A deployer is any organisation using an AI system under its authority in a professional capacity. A company whose staff use a ready-made chatbot for work is therefore covered, not only companies that build AI.
The Digital Omnibus on AI, in force since 27 July 2026, amended this article (Regulation (EU) 2026/1744). The law now requires measures that support AI literacy. It does not require companies to guarantee a particular level of literacy for each employee. The full timeline is in what the EU AI Act requires.
This is not a reason to drop the topic. The obligation still applies, and a company that uses AI at work but has done nothing to help staff understand it is not meeting it. In practice, that means measures proportionate to the company’s size, its tools and the risks involved, together with a record of what was done.
A role-to-knowledge matrix
Recommendation. Match the depth of training to each group’s role. One person can belong to several groups, and the basics come first: what AI can and cannot do, and why its output needs checking.
| Group | What they need to know | Suggested format |
|---|---|---|
| All staff | Where the company uses AI; that output can be wrong or biased; the company’s AI rules; which data never goes into AI tools; how to report a problem | Session of 30–60 minutes, repeated for new joiners |
| Daily AI users | How to write prompts and check output; the limits of the tools they use; when a person must review; when AI use must be disclosed | Hands-on session with the actual tools |
| Use-case owners | How to assess and approve a use case; vendor terms; human oversight; monitoring and incidents; the company’s role under the AI Act | Workshop based on internal procedures |
| Developers | Data quality and bias testing; security risks such as prompt injection; logging and documentation; the requirements that apply to what they build | Technical training alongside secure development practice |
| Contractors and freelancers | The same basics as all staff, plus the rules for the tools they use or operate on the company’s behalf | Short briefing at onboarding, with the AI use policy |
| Leadership | Where AI creates value and risk; legal obligations and their dates; who is accountable; when to pause or stop a use case | Short briefing, repeated when rules or tools change |
Build the content around what people use today, taken from your AI inventory, rather than a generic course. Training on the real tools and on your own AI use policy is shorter and easier to apply.
Keep a record
Recommendation. For each session, note the date, the topic, who delivered it, who attended and where the materials are stored. Update the record when people join, change role or start using a new tool. Keep it with your other AI records, so that you can show which measures were taken and when.
A record does not replace the training itself. A signed attendance sheet for a session nobody remembers shows little; short, repeated sessions tied to real tasks do more to build AI literacy than a single long course.
Next step: run a short introductory session for all staff, and for contractors who use AI on your behalf, on the tools you use and your most important rules, and record who attended.
Sources and further reading
- Regulation (EU) 2024/1689 (AI Act) — Articles 3 and 4
- Regulation (EU) 2026/1744 (Digital Omnibus on AI)
This article is for general information and is not legal advice.
AI Horizon Conference
The AI Horizon Conference returns to Lisbon, once again bringing together entrepreneurs, investors and industry leaders to discuss the future of AI.
