What Article 4 requires
Article 4 of Regulation (EU) 2024/1689 — the EU AI Act — is remarkably short: providers and deployers of AI systems must take measures to ensure, to their best extent, a sufficient level of AI literacy among their staff and other persons dealing with AI systems on their behalf.
“Sufficient” is not a fixed standard. It depends on context: the technical knowledge, experience and training of the staff, the context in which the AI systems are used, and the persons affected by their use. A clerk who occasionally drafts texts with an AI assistant needs a different level than a team using AI in candidate selection.
The direction of the rule matters: the goal is not to turn everyone into a technical expert. AI literacy in the sense of the regulation means using AI systems knowledgeably and being aware of their opportunities, risks and possible harms — which is exactly what this course teaches.
Who is covered: practically every company
The duty applies to “providers” (those who develop AI systems or offer them under their own name) and to “deployers” — in the regulation's vocabulary, any organization using an AI system under its own authority in a professional context. Running an AI chatbot for customer enquiries, drafting texts with an AI assistant, pre-sorting applications with AI support: all of this makes a company a deployer.
This means: even companies that “only” use common tools such as ChatGPT or Microsoft Copilot in daily work are covered. Company size is irrelevant — a five-person trade business is covered just like a corporation. Only purely private use is out of scope.
The duty has applied since 2 February 2025 — together with the bans on certain AI practices, it was among the first parts of the AI Act to become applicable. There is no remaining transition period: whoever uses AI today must ensure their staff's literacy today.
What happens if you don't comply
The AI Act's fine catalogue (Article 99) does not contain a separate fine solely for missing training — that should be said honestly. But concluding that the duty has no consequences would be a mistake.
First, Article 4 is applicable law whose observance national supervisory authorities can review and enforce as part of market surveillance. Second, missing AI literacy spills into other obligations: many AI Act requirements — such as human oversight of high-risk systems — simply presuppose trained staff. Third, in case of damage, organizational fault matters: if an employee causes harm with an AI tool — say, by passing on an unchecked, incorrect statement — the question arises whether the company trained its people adequately. A documented training is then a robust defence.
In short: the training duty is less a fine threat than a due-diligence and evidence question — which is exactly why documentation is part of it.
The answer: train and document
The European Commission has clarified that there is no mandatory certificate programme — companies may decide themselves how to build the competence. What is expected is an approach that matches the actual use of AI: a baseline training for everyone working with AI, and deeper training where use is riskier.
A simple package works well: first, get an overview of which AI tools are actually used in the company; second, have all affected staff complete a baseline training like this one; third, document participation and content (who, when, what); fourth, keep it up to date as new tools or use cases arrive.
That is exactly what this course is built for: the following lessons cover how AI works (lesson 2), productive use (lesson 3), safe handling of risks (lesson 4) and the overall legal framework (lesson 5). Completing all lessons including the knowledge checks covers the core content of an Article 4 baseline training.