Since 2 February 2025, the EU AI Act requires a sufficient level of AI competence in your operations. What that means in practice, who it affects and how you put it into place — all on this page, without scaremongering about fines.
The AI Act (Regulation (EU) 2024/1689) has been in force since August 2024. Article 4 on AI competence applies since 2 February 2025 — it affects SMEs that merely use AI, too.
Providers and deployers must ensure a sufficient level of AI competence among their staff.
The competence duty has applied since 2 February 2025 — regardless of the risk class.
Anyone using AI in their operations (e.g. ChatGPT, copilots) already counts as a deployer.
Article 4 obliges providers and deployers to take measures, to their best extent, to ensure that their staff — and other persons dealing with AI on their behalf — have a sufficient level of AI competence.
“Sufficient” depends on the context: it takes account of the technical knowledge, experience and education of those involved, the specific area of use, and the groups of people the AI is applied to. So there is no fixed number of hours — but a traceable, appropriate build-up of knowledge.
The duty applies to both roles. For most technical SMEs the second one is decisive — merely using AI is already enough.
Develops an AI system (or has it developed) and places it on the market or puts it into service under its own name or brand.
Uses an AI system under its own authority in its operations — for example ChatGPT, Microsoft Copilot or AI features in CAD/ERP software.
The AI Act enters into application in stages. For AI competence, the deadline has already been reached.
Regulation (EU) 2024/1689 enters into force — the start of staggered application.
The competence duty applies — regardless of the risk class. Prohibited AI practices are banned.
Transparency and governance obligations for GPAI models take effect.
The bulk of the obligations for high-risk applications become effective.
Extended deadline for AI in products that are already subject to other EU product regulation.
A pragmatic path — understandable, documentable, without a bureaucratic monster.
Which AI tools are in use — officially and unofficially? Who uses what, and for what?
Provider or deployer? Which areas of use, which data, which risk class?
In-house and on real tasks: possibilities, limits, prompting, source criticism, data protection.
Clear rules: approved tools, how to handle company data, what may go to the cloud and what may not.
Record trainings, rules and approved tools traceably — and update them regularly.
Yes. Article 4 does not depend on company size. As soon as AI is used in your operations, the competence duty applies — including in small plant-construction or manufacturing businesses.
A good start, but competence is ongoing: tools and use cases change. A recurring refresher plus a well-maintained AI policy makes sense.
It should be traceable which trainings took place, which rules apply and which tools are approved. No bureaucratic monster — but reliable.
We train your team in-house on real tasks and accompany the implementation, including AI policy and documentation — ready for the AI-Act duty.
This page provides a plain-language overview and is not legal advice. The wording of Regulation (EU) 2024/1689 prevails; consult qualified advice for your specific case.
We train your team in-house, hands-on and data-protection-compliant — and accompany a clean implementation. Request the course overview.