AI Literacy
What is AI Literacy
AI Literacy is the ability to understand AI systems — what they can do, what they cannot, what risks they carry, and how to use them safely. In the context of EU regulations, it is an obligation, not a choice.
AI Act Art. 4 — AI Literacy Obligation
Article 4 of the AI Act (effective since 2 February 2025) stipulates:
Providers and deployers of AI systems shall take measures to ensure a sufficient level of AI literacy of their staff and other persons dealing with the operation and use of AI systems on their behalf.
What this means in practice
| Aspect | Requirement |
|---|---|
| Who | Everyone who works with AI systems (not just IT) |
| What | Understanding how AI works, its limitations and risks |
| When | Obligation effective since 2 February 2025 |
| How | Training proportionate to the role, context and risk |
| Demonstrability | Organizations must be able to demonstrate the measures taken |
What “proportionate” means
The AI Act does not specify the exact scope of training — it requires that training be proportionate to:
- Role — a developer needs a different level than a receptionist
- Context — healthcare requires a different focus than marketing
- Risk — working with high-risk AI systems requires deeper training
Three levels of AI literacy
| Level | For whom | What it covers |
|---|---|---|
| Basic | All employees | What AI is, what the rules are, how to use it safely |
| Advanced | AI tool users | Prompt engineering, tool selection, critical evaluation |
| Expert | AI practitioners, management | Architecture, governance, risk assessment |
Basic vs. advanced literacy
Basic literacy (L1 — Align) satisfies the regulatory obligation:
- The employee knows what AI is and how it manifests in the organization
- Knows the company AI policy and rules
- Has completed training and knows how to report an AI incident
Advanced literacy (L2 — Transform) delivers real value:
- The employee can effectively use AI in their work
- Masters prompt engineering and knows relevant tools
- Critically evaluates AI outputs
How to ensure AI literacy
- Map — identify who in your organization works with AI systems
- Classify — determine what level of literacy each role requires
- Train — deliver training proportionate to the roles
- Document — keep records of training completed
- Update — literacy is not a one-time event, AI evolves continuously