EU joins the Council of Europe’s Convention on Artificial Intelligence

On 5 September 2024, in Vilnius, the European Union acceded to the Council of Europe's Convention on Artificial Intelligence, making the Convention binding for all EU Member States. The Convention was adopted by the Committee of Ministers of the Council of Europe in Strasbourg on 17 May 2024 and the Council of the European Union decided on 28 August to be among the first parties to sign it.

The importance of the Convention lies in the fact that it is the first legally binding international instrument to regulate artificial intelligence. The Convention focuses on three main areas: human rights, respect for democracy and the rule of law, and a framework for the whole life cycle of AI in terms of innovation and risk management.

The EU has already been actively involved in the preparation of the Convention, so the Convention is fully in line with EU legislation, especially the also landmark EU AI Act. The convention seeks to ensure that the risks posed by AI to human rights, democracy and the rule of law can be addressed at international level through a number of mechanisms. These mechanisms include a risk-based approach to AI, the development of principles for reliable AI (transparency, reliability, security), documentation obligations and oversight mechanisms to monitor AI activities.

The importance of the innovations and rules cannot be underestimated, but the Convention has also been criticized, as the scope of the obligations is so broad that the effective enforceability of the Convention is questionable.

In connection with the signing of the Convention by the EU, it is worth mentioning Hungary, or more precisely the compromise proposal endorsed by the Hungarian Presidency. The proposal is for a so-called two-step approach, whereby the Convention would first be signed by the EU on behalf of the Member States, and then, if necessary, in a second step, the EU could empower the Member States to sign the Convention in the general interest of the EU.

The Convention has been signed by the European Union, the United States, the United Kingdom, Norway, Iceland and Israel, among others, so the regulation of activities in the life cycle of AI systems can be globally harmonized.