AI glossary for businesses
AI literacy
In short
AI literacy is the knowledge and skills people need to use AI tools sensibly: what they do well, where they fail and what risks they bring. Under Article 4 of the EU AI Act, organisations that provide or use AI systems must take measures to ensure sufficient AI literacy among their staff, a duty in force since 2 February 2025.
How AI literacy works
AI literacy is not about learning to code. It is closer to a driving licence for AI tools. You do not need to know how the engine works, but you must know the rules of the road, the blind spots and when to brake.
For most staff, that means a few practical abilities:
knowing what tools like ChatGPT, Microsoft 365 Copilot or Google Gemini do well, and where they make things up;
writing clear prompts and checking the output before using it;
knowing which data must never go into an unapproved AI tool;
knowing who to ask when something looks wrong.
Article 4 of the EU AI Act asks providers and deployers of AI systems to take measures, to their best extent, to ensure a sufficient level of AI literacy for staff and other people who operate AI on their behalf. The level should reflect their technical knowledge, experience, education and training, and the context in which the AI is used. The text does not prescribe a specific course or certificate.
In practice for a small business
Take a retail chain with 40 employees in three Greek cities. Store managers use ChatGPT for staff rotas and announcements. The marketing team uses Canva and Claude for campaigns. A sensible approach could be a short session for everyone, a one-page AI usage policy, and a deeper session for the marketing team, who use AI daily. Keep a simple record of who was trained, when and on what. That record helps show you took the duty seriously. The article on preparing your team for AI adoption has more ideas.
Why it matters and what to watch out for
It already applies. The duty has applied since 2 February 2025, also to companies that only use off-the-shelf tools.
One size does not fit all. A receptionist and a data analyst need different depth. Match training to the role and the risk.
It reduces real risks. Trained staff are less likely to paste customer data into the wrong tool or publish an AI error. This also supports your GDPR obligations.
Refresh it. Tools change fast. Plan a short update at least once a year.
Not legal advice. For what counts as "sufficient" in your case, check with your legal advisor.