EU AI Regulation & Training Obligation: What Lies Ahead for Enterprises – and How to Prepare Optimally

With the entry into force of the central provision of the EU-KI-Verordnung (also "EU AI Act") on 02.02.2025, a new era of regulation in the handling of artificial intelligence began for enterprises. The requirement: enterprises must ensure that all employees working with these systems are appropriately trained ("AI competency").

The objective of the EU-KI-Verordnung is to create trust and safety in the deployment of AI systems, in particular for so-called high-risk AI applications, such as those used in personnel recruitment, medical diagnostics or in the financial sector.

The training obligation applies to all enterprises and organisations that high-risk AI systems within the meaning of the Regulation either develop, provide or use. This affects in particular:

• Manufacturers of AI systems active in sensitive areas such as health, education, justice, critical infrastructure or employment mediation;
• User enterprises that deploy such systems in their ongoing business – regardless of enterprise size;
• Importers and distributors that place high-risk AI systems on the market in the EU or put them into circulation.

Small and medium-sized enterprises (SMEs) and start-ups are also affected if they use or offer high-risk AI systems. Whilst the Regulation does contain eased requirements for SMEs, the obligation to provide training nevertheless remains, as it is central to safe application.

The Regulation provides that all persons entrusted with the operation, monitoring or interpretation of a high-risk AI system must possess the necessary competencies. These trainings are intended to ensure that employees understand how the system functions, are aware of possible risks (such as discrimination or erroneous results) and know how to handle unexpected behaviour of the AI. Aspects such as data protection, transparency and responsible handling of training and output data must also be covered in training. The EU-KI-Verordnung does not prescribe fixed contents or uniform training formats; rather, it obligates enterprises to provide tailor-made trainings on the basis of the specific systems deployed and their risks. Training must therefore always be appropriate to the system, the deployment context and the role of the person being trained.

The following applies: trainings must be updated on a regular basis, in particular if the AI system is further developed or new legal requirements come into force.

AI applications that are not classified as high-risk AI (for example, simple chatbots, recommendation algorithms, marketing tools) are not subject to the formal training obligation under the Regulation. Nevertheless, for reasons of corporate responsibility and compliance, it is advisable to conduct at least basic trainings on the safe and data protection-compliant use of such applications.

Non-compliance with training obligations can result in significant legal consequences. The Regulation provides for substantial fines – in the event of violations of operator obligations, up to EUR 10 million or 2 % of worldwide annual turnover. In addition, usage bans for systems not deployed in a regulation-compliant manner and potential civil liability are threatened if untrained personnel cause damage. Public disclosure of violations can also result in considerable reputational risk.

Recommended actions for enterprises in dealing with the EU AI Regulation

In order to meet the new requirements, enterprises should take the following measures in good time:

1. Conduct a stocktake: Check whether and which high-risk AI systems are in use in the enterprise.
2. Identify training needs: Determine which employees must be trained – also across departments.
3. Develop training concepts: Create training content tailored to the target groups and adapted to the respective use.
4. Ensure documentation: Record trainings and participation systematically in order to be prepared in the event of an inspection.
5. Clarify responsibilities: Designate internal points of contact for compliance with AI regulations.
6. Establish continuous further training: Integrate trainings on AI risks and ethical use into your regular training offer.

Implementation of these obligations is not merely a legal necessity, but also an important contribution to the responsible use of artificial intelligence in the enterprise. Those who act early secure themselves a compliance advantage – and at the same time strengthen the trust of employees and customers.

Should you have further questions regarding legally compliant implementation of the EU AI Regulation or for trainings in your enterprise, the lawyers at AVANTCORE Rechtsanwälte are on your side in this European matter not only in the heart of Stuttgart, but also across borders with profound expertise in IT-Recht and Arbeitsrecht.