EU AI Act · Article 4
One baseline for everyone.Different depth for different reach.
Article 4 asks providers and deployers to take measures for sufficient AI literacy. Sufficiency depends on people, knowledge, context, and the systems they use — not one generic certificate.
25 minutes · free trial up to five people · no payment card
before training
Can you evidence the programme today?
This check is not a legal assessment. It shows whether the basic programme controls are visible and documentable before you choose more content.
five-question readiness check
Answer from the evidence you could produce today, not from policy intent.
01We know which AI systems and agents staff use, including unsanctioned tools.
02Each person is mapped to training depth based on what AI can reach or affect.
03Our programme covers opportunities, risks, possible harm, and safe action.
04We record content version, completion, assessment, and attestations.
05We refresh competence and practise incidents rather than relying on one course.
Anonymous and ephemeral — these answers are not stored and are not a compliance determination.
verified free baseline
What each learner completes
8 min
What Article 4 asks
The duty, who it covers, and why sufficiency depends on context.
8 min
Opportunities, risks, harm
Classify useful capability, unsafe conditions, and possible consequences.
9 min
Your baseline actions
Practise three decisions, pass a five-question quiz, and sign the participation record.
- Verified account and named participation record
- Course and lesson versions pinned in the evidence trail
- Server-derived quiz result and attested wording
- Clear statement that participation is not accreditation or legal advice
after the baseline
Training follows what an agent can do on your behalf.
All staff continue into agent-safety foundations. Operators add delegation and access gates. Builders and deployers add capability, logging, vendors, and incident runbooks. Governance roles add liability, evidence duties, and tabletop practice.