This January there was good progress in the Council of Europe’s negotiations for an international convention on artificial intelligence (AI). The EU is committed to engage in this process and work towards an agreement on binding common principles and rules for AI.

 Council of Europe convention on AI and EU AI act

In 2022, the Council of Europe set up the Committee on Artificial Intelligence (CAI) to start a negotiation process for a legal instrument on the development, design, and application of AI based on the Council of Europe standards for human rights, democracy, and the rule of law. The instrument will be of a legally binding nature (convention or framework convention) and will be open to Council of Europe members, the European Union and third countries who share the same values. The current timeline for the CAI negotiations aims to an agreement on the text of the convention by November 2023.

Following a Council decision from November 2022, the European Commission is now authorised to negotiate the convention on AI on behalf of the European Union. Our commitment does not just go hand in hand with EU’s priority to work together with like-minded countries for the promotion of human-centric and trustworthy AI but it is also natural as the EU heads towards the adoption of its own legal framework on AI, proposed by the European Commission in April 2021.

By the end of 2023, the EU is expected to adopt harmonised rules on AI that would be directly applicable in all 27 EU Member States within the EU internal market. The so called “AI act” will address two main objectives: (i) protecting from potential risks that AI can pose to people’s safety and fundamental rights and (ii) creating a clear regulatory environment where innovative and trustworthy AI applications can easily access the market.

To achieve these two objectives, the AI act proposes the classification of risks created by the use of AI in different levels and specific requirements based on that classification. Thanks to a future-proof approach and its complementarity to existing legislation, the new rules are expected to adapt to technological change and give space to innovation.

While the AI act is world’s first regulation in this field to be directly applicable in all 27 EU member States, the future Council of Europe convention on AI will have to be implemented in the legal order of the parties through their domestic legislation. This makes the EU AI act automatically an instrument through which the disposals of the Council of Europe convention on AI can be integrated in the EU legal order, if the EU joins the convention.

Status and next steps of negotiations

The EU has reiterated its strong support for the work done by the Council of Europe to promote human rights, including in the field of AI. Key orientations for the EU will be to follow a risk-based and future-proof approach that leads to proportionate, effective, and clear principles for AI. The final goal is to formulate a flexible framework convention that avoids overlaps and provides an added value to other relevant instruments and conventions and which is consistent with the AI Act.

The CAI meeting in January made significant progress in defining the working methods of the convention’s roadmap and clarifying the legal nature for the risk and impact assessment methodology. The group also held a first discussion on the final provisions of the convention.

The next plenary will be held on 1-3 February to discuss the changes proposed by CAI’s drafting group to the final provisions and the follow-up and cooperation mechanism of the convention as well as its preamble.

You can find more updates from the negotiation process on CAI’s webpage.