The Council of Europe on Thursday opened for signature the world’s first legally-binding international treaty on artificial intelligence (AI). Announced at a conference in Vilnius, Lithuania, the historic treaty establishes new international standards to ensure that AI systems are consistent with human rights, democratic values, and the rule of law.
Officially titled the Framework Treaty on Artificial Intelligence, Human Rights, Democracy and the Rule of Law (CETS No. 225), it marks a pivotal milestone in global AI regulation. It is the first international agreement on AI governance and has been signed by the European Union (EU), the UK, the US, Israel, Andorra, Georgia, Iceland, Norway, Moldova and San Marino. Adopted by the Committee of Ministers of the Council of Europe on 17 May 2024, the Treaty establishes a comprehensive legal framework covering the entire lifecycle of AI systems, from design and development to deployment and decommissioning. It uses a technology-neutral approach designed to adapt as AI evolves, while promoting responsible innovation while addressing potential risks.
The Treaty is the first legally binding global framework on AI that is harmonised with EU law, including the EU AI Law, the world’s first far-reaching AI regulation. The EU, through the Commission and its Member States, played an active role in the negotiating process and contributed significantly to the development of the Treaty. The Treaty is in line with the principles outlined in the EU AI Law and other EU regulations and incorporates fundamental elements such as:
Focus on human-centric AI aligned with human rights, democracy and the rule of law Risk-based approach Key principles of Trustworthy AI (e.g. transparency, robustness, safety, data governance and protection) Transparency of AI-generated content and interactions with AI systems Enhanced documentation, accountability and remedies Supporting safe innovation through regulatory sandboxes Risk management obligations Documentation obligations Oversight mechanisms for AI activity
The Convention will enter into force three months after it has been ratified by at least five signatory countries, including three that are Council of Europe member states. This provision also allows time for the Convention to be effectively implemented and widely adopted. Negotiated by the Council of Europe, 46 Member States, the EU, and 11 non-member countries contributed to the Convention, including Argentina, Australia, Canada, Costa Rica, the Holy See, Israel, Japan, Mexico, Peru, the United States, and Uruguay. Representatives of the private sector, civil society, and academia also participated as observers.
Secretary-General of the Council of Europe, Marija Pejčinović Burić, emphasised the importance of the Convention, saying:
We must ensure that the rise of AI upholds, rather than undermines, our standards. The Framework Convention is designed to ensure just that. It is a strong, balanced document, the result of being drafted with an open and inclusive approach, ensuring that it benefits from multiple expert perspectives. The Framework Convention is an open treaty with potentially global implications. I hope that these will be the first of many signatures, with ratifications to follow soon after, so that the Convention can enter into force as soon as possible..
Additionally, parties to the treaty must guarantee legal remedies for victims of AI-related human rights violations and provide procedural safeguards, such as notifying individuals interacting with AI systems. The treaty also requires parties to prevent AI from undermining democratic institutions and processes, including separation of powers, judicial independence, and access to justice. The treaty does not apply to national security activities, but requires such activities to comply with international law and democratic principles. National defense and research and development activities are also exempt, except where AI testing may affect human rights, democracy, or the rule of law.
