Official name: Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law
Citation: Council of Europe Treaty Series No. 225, opened for signature on 5 September 2024
Year / status: Open for signature; yet to enter into force
What it does
The Council of Europe, the Strasbourg-based human rights organisation of 46 European states, negotiated this framework convention with its members, with observer and other countries, with the European Union, and with representatives of civil society, academia and industry. It was opened for signature in Vilnius on 5 September 2024. The Council of Europe calls it the first legally binding international treaty on artificial intelligence. It requires that activities across the whole lifecycle of an AI system respect human rights, democracy and the rule of law.
Where it applies / status
The convention applies in full to public authorities and to private companies acting on their behalf. For other private actors, each country must address the risks in line with the treaty’s aims and declare how it will do so: by applying the treaty’s duties directly or by other suitable measures. Countries may leave out activities that protect national security, provided they respect international law, and national defence falls outside the treaty altogether. Research and development on systems before they are made available for use is also left out, unless testing could interfere with human rights, democracy or the rule of law. The treaty enters into force on the first day of the month after three months have passed from the date when five signatories, at least three of them Council of Europe member states, have ratified it.
Documented benefits or intended effects
Countries that ratify the convention take on legal duties in their own law. The treaty sets shared principles: human dignity and individual autonomy, equality and non-discrimination, privacy and personal data protection, transparency and oversight, accountability and responsibility, reliability, and safe innovation. It also calls for accessible and effective remedies when an AI system violates human rights, for procedural safeguards, for telling people when they are interacting with an AI system, and for measures to identify, assess and reduce risks. The wording is technology-neutral, so it can also cover systems built in the future.
Caveats
On 29 September 2026 the Council of Europe Treaty Office listed 21 signatories: 15 Council of Europe member states, among them the United Kingdom, Norway, Switzerland and Ukraine; five other countries, namely Canada, Israel, Japan, the United States and Uruguay; and the European Union. Only the European Union had ratified, on 15 May 2026, so the treaty had yet to enter into force. A signature signals intent; legal duties start with ratification and entry into force. Each country decides how the rules reach private companies, so protection can differ from one country to the next. The Treaty Office chart gives the current list of signatures and ratifications.
Sources
Council of Europe treaty page, treaty text, chart of signatures and ratifications, and the news of ratification by the European Union.
- Council of Europe: The Framework Convention on Artificial Intelligence
- Council of Europe: Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law (Treaty Series No. 225)
- Council of Europe Treaty Office: Chart of signatures and ratifications of Treaty 225
- Council of Europe: European Union ratifies the Framework Convention on Artificial Intelligence
- Wikimedia Commons: Council of Europe Palais de l'Europe aerial view (photo)
