First bans of the European Artificial Intelligence Regulation (RIA) enter into force

European Regulation on Artificial Intelligence (RIA)

On 3 February 2025, the first obligations and prohibitions of the European Regulation on Artificial Intelligence (RIA)created in May 2024 to mitigate the risks of artificial intelligence and protect the rights and freedoms of all people in the world. European Union. This regulation establishes a legal framework for the development, marketing and use of AI systems within the EU.

Who is affected by the legislation

The European Regulation on Artificial Intelligence (RIA) It concerns all providers of AI systems, regardless of whether they are established inside or outside the European Union, and those responsible for the deployment of AI systems established in the Union.

All companies and organisations must ensure that their staff are properly trained in the use of these tools and comply with the regulations, otherwise, in the event of non-compliance, they may face severe penalties of up to 35 million or the 7 % of annual turnover.

Categories of infringements and associated penalties

The European Artificial Intelligence Regulation (EAR) establishes a rigorous sanctioning regime to ensure compliance with its provisions and to protect the fundamental rights of European citizens. Penalties are structured according to the seriousness of infringements and can reach significant amounts.

Categories of infringements and associated penalties

  1. Very serious infringements:
    • DescriptionThey include the development, marketing or use of prohibited AI systems, such as those that manipulate human behaviour through subliminal techniques or that exploit specific vulnerabilities of vulnerable groups.
    • SanctionFines of up to EUR 35 million or 7% of the undertaking's total annual turnover, whichever is higher.
  2. Serious infringements:
    • Description: These refer to non-compliance with obligations by suppliers, authorised representatives, importers, distributors or those responsible for the deployment of high-risk IA systems. This includes failure to ensure compliance of systems with established requirements, such as risk management, technical documentation or human supervision.
    • Sanction15 million or 3% of the undertaking's annual aggregate turnover, whichever is higher.
  3. Minor infringements:
    • DescriptionThey cover the provision of incorrect, incomplete or misleading information to notified bodies or national competent authorities in response to an application.
    • Sanction7.5 million or 1% of the undertaking's annual aggregate turnover, whichever is higher.

Criteria for the imposition of sanctions

Sanctions must be effective, proportionate and dissuasive. In determining the level of fines, factors such as the gravity and duration of the infringement, the degree of responsibility, the measures taken to mitigate the damage, cooperation with the authorities and any relevant previous infringements are considered.

Responsibility of Member States

Each EU Member State is responsible for establishing and implementing its own penalty regime for breaches of the RIA. These penalties may include, in addition to administrative fines, non-financial measures such as warnings or the suspension of the placing on the market of non-compliant AI systems. Member States must notify the provisions on penalties to the European Commission by 2 August 2025.

Next phases of implementation

The European Regulation on Artificial Intelligence (RIA) provides for a staggered implementation of its provisions:

  • 2 August 2025The rules on AI models for general use will enter into force.
  • 2 August 2026Most of the provisions of the regulation will be applicable as of this date.
  • 2 August 2027Obligations associated with high-risk AI systems related to regulated products.

During this transitional period, the European Commission has launched the AI Pact, an initiative that invites developers to voluntarily adopt key RIA obligations ahead of the legal deadlines.

Conclusion

The sanctioning regime of the European Regulation on Artificial Intelligence underlines the commitment of the European Union with the development and responsible use of the IA. Companies and organisations must ensure that they comply with the obligations set out to avoid significant penalties and contribute to safe and ethical AI in Europe.

*AI-generated text from the sources: