Council of Europe Framework Convention on AI
The Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law (Council of Europe; opened for signature 5 Sept 2024, Vilnius) is the first legally binding international treaty on AI — a different instrument class from the eu-ai-act, and a key piece of the “whose standards win?” question. Source: Wikipedia.
What it is
- Binding, but principles-based: seven core principles (human rights, transparency, accountability, non-discrimination, rule-of-law) across 8 chapters / 26 articles, with a risk-management structure — not product classifications.
- Broad signatory base beyond Europe: “more than 50 countries,” incl. the EU, US, UK, Canada, Japan, Switzerland, Ukraine — extending past Council-of-Europe membership.
Treaty vs. regulation (the key distinction)
The contrast this spoke cares about: the eu-ai-act is a regulation with risk-tiered product rules enforced inside the EU; the Framework Convention is a treaty binding signatory states to principles for AI affecting human rights/democracy. Two binding routes to risk-based-regulation — one regulates products, the other commits governments — both sitting above the voluntary oecd-ai-principles/iso-iec-42001 tier. A useful counterexample to the framing that Israel’s sectoral soft law vs. the EU act are the only two models.
Related
ai-governance · eu-ai-act · oecd-ai-principles · risk-based-regulation · responsible-innovation