On 21 April 2021, the Commission tabled a proposal for a regulation laying down harmonised rules on artificial intelligence (AI). Following the Commission’s White Paper on AI from 2020 (→ eucrim 1/2020, 8-9) and in a new step aiming to turn Europe into the global hub for trustworthy AI, the proposal strives to balance the numerous risks and benefits the use of AI can provide.

In order to implement the second objective of the White Paper addressing the risks associated with using AI in certain contexts, the Commission is proposing a legal framework that will enable the benefits the use of AI has to offer to be reaped, while simultaneously upholding the EU’s values and fundamental rights. The Regulation on AI pursues four objectives:

  • To ensure that AI systems placed on the EU market are safe and in line with existing EU law on fundamental rights and values;
  • To ensure legal certainty when facilitating investment in and innovation into AI;
  • To enhance governance and effective enforcement of the existing law on fundamental rights and safety requirements applicable to AI systems;
  • To facilitate the development of a single market for lawful, safe, and trustworthy AI applications and to prevent market fragmentation.

The direct application of the new rules across all Member States will be based on a future-proof definition of AI that is based on a risk-based approach, going from unacceptable-risk to minimal-risk AI systems:

  • AI systems that are considered an unacceptable risk, i.e., a clear safety threat to people and “contravening Union values,” will be banned. The Commissions sees such unacceptable risks in AI systems that allow “social scoring” by governments.
  • AI systems that are identified as high-risk, i.e., posing significant risks to the health and safety or fundamental rights of persons, will be put under strict obligations before being placed on the European market. This includes AI systems that are being used, for example, in critical infrastructures, migration, asylum and border control management, or in the administration of justice and democratic processes. The Commission stresses that all remote biometric identification is considered high-risk and will therefore be subject to specific restrictions and safeguards.
  • Limited-risk systems, i.e., chatbots, will be subject to minimum transparency obligations.
  • Minimal-risk AI systems that represent only a minimal or no risk for the rights and/or safety of citizens, i.e., spam filters, are not subject to the proposed Regulation.

On the governance side, the proposal establishes a European Artificial Intelligence Board at the EU level, which is to be tasked with contributing to effective cooperation between the national supervisory authorities and the Commission and with providing advice and expertise to the Commission.

In parallel to the first worldwide initiative to set up a legislative framework on AI, the Commission presented the following documents:

A proposal for a regulation on machinery products, which ensures that the new generation of machinery guarantees the safety of users and consumers and encourages innovation.