EU AI Act – EU Artificial Intelligence Regulation

From 2 August 2025, the obligations for general-purpose AI models and the governance framework will also enter into force.
2 August 2026 is the general entry into force of the regulation, which in practice is the main compliance deadline: companies developing or using high-risk AI systems must comply with the obligations under the regulation by this date.

An extended deadline for certain embedded AI systems will last until 2 August 2027. For high-risk integrated AI systems that are part of existing products subject to other EU product regulations (e.g. AI components in medical devices or vehicles), a one-year grace period has been set.

Although the regulation primarily applies to service providers, distributors and importers, the provisions affecting users of AI-based systems cannot be ignored, given that in the event of non-compliance with the obligations arising from the regulation, the competent authority may impose, inter alia, the following fines:

In the event of non-compliance with the prohibition of prohibited AI practices, an administrative fine of up to EUR 35,000,000 or, in the case of an undertaking, up to 7% of its total global annual turnover in the preceding financial year, whichever is higher;

If the AI-system does not comply with the requirements for economic operators or notified bodies, in the case of users, set out in Articles 26 and 50 of the Regulation, administrative fines of up to EUR 15 000 000 or, in the case of an undertaking, up to 3% of its total global annual turnover in the preceding financial year, whichever is the higher, shall be imposed.

Under the Regulation, a user is any natural or legal person who uses an AI-system under its control in the context of a professional activity, including employers who use artificial intelligence for purposes such as recruitment, performance assessment, work organisation or monitoring.

Users of AI-systems shall ensure that their staff and any other person involved in the operation and use of AI-systems on their behalf have acquired an appropriate level of AI competence.