AI literacy is becoming an important part of responsible AI development and deployment. As organisations integrate AI into their products and workflows, employees need to understand how these systems work, where their limitations lie and what risks they may introduce.
Article 4 of the EU AI Act addresses this issue by requiring providers and deployers of AI systems to take measures to ensure, to their best extent, a sufficient level of AI literacy among their staff and other people dealing with AI systems on their behalf.
For engineering teams, compliance teams and AI product managers, this raises a practical question: how should an organisation organise its AI literacy activities and document its approach?
1. Start with an AI system inventory
Before planning training, identify the AI systems used or developed across the organisation.
- For each system, consider recording:
- System name and intended purpose
- Business or technical owner
- Teams and people interacting with the system
- Relevant operational risks
- Existing guidance and training needs
An AI inventory helps teams connect AI literacy activities with actual use cases rather than relying on generic training.
2. Map AI literacy needs to different roles
A developer building an AI-powered feature and an employee using a generative AI assistant have different responsibilities.
Developers may need guidance on system limitations, testing, data handling and risk management. Business users may need practical instructions on output verification, confidentiality and appropriate use.
An effective AI literacy program takes these differences into account. The right measures depend on the person's knowledge, experience, education, training and the context in which the AI system is used.
3. Maintain useful AI literacy documentation
Article 4 does not prescribe one universal training template or require a specific AI literacy certificate. However, maintaining internal records can help an organisation understand and review its approach.
Consider documenting:
- The learning activities or guidance selected
- The intended audience
- The date and purpose of training activities
- Supporting materials or evidence, where appropriate
- Follow-up actions and review decisions
These are practical documentation recommendations, not a mandatory statutory checklist.
4. Review the program when AI use changes
AI systems and business workflows evolve. A new AI application, a change in intended use or a shift in employee responsibilities may create new learning needs.
Reviewing AI literacy measures when relevant changes occur helps teams keep their guidance useful. Organisations can also consider periodic reviews to identify gaps and coordinate activities across departments.
5. Connect AI literacy with broader AI governance
AI literacy is one part of a wider governance process. Organisations may also need to manage AI inventories, risk classification, technical documentation, human oversight and other applicable obligations under the EU AI Act.
AI compliance software can help organise related information and workflows, but software alone does not establish compliance. Teams still need to assess their circumstances and select appropriate measures.
Final thoughts
AI literacy documentation should reflect how an organisation approaches learning, who interacts with its AI systems and why particular measures have been selected.
A clear AI inventory, role-based learning activities and well-organised internal records can give AI teams a practical foundation for managing this work.
For a more detailed breakdown, read the full guide: AI Literacy Documentation: What Should Companies Record?
AnnexOps supports AI compliance and governance operations by helping organisations structure relevant workflows and documentation.
Top comments (0)