Make learning relevant to each role
The EU lists AI literacy among the AI Act provisions applicable since February 2025. Appropriate measures depend on knowledge, experience and deployment context, among other factors. This does not create a universal requirement for a particular course or certificate.
Training works particularly well with examples from actual use. Summarising meetings requires different knowledge from reviewing AI outputs for employment-related decisions. Tasks and risks therefore shape the learning content.
Four topics for the first learning session
Approved examples from everyday work make typical errors understandable. Exercises in source review, missing information and permitted data types build practical competence. This includes distinguishing a draft from an approved action.
- What a language model can do and where it fails
- Which data may be used in which tool
- How sources and numbers are checked
- When a person must take over or stop
Make learning visible in practice
Record learning objectives, examples discussed and ownership. Repeat exercises when new features or data types are introduced. AI Link can hold agreed working practices in team knowledge, but a shared document does not replace practice.
Sources & further reading
Original research, provider documentation and further reading on the concepts and figures discussed here.