The First AI Treaty as Expressive Hard Law: Signature, Supervision, and the Space Between
This essay argues that the Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law, the first binding international treaty on artificial intelligence (AI), is best understood as expressive hard law. The Convention opened for signature in September 2024 and drew broad signatory support, yet as of September 2026, no country has ratified it, it remains short of entry into force, and its Conference of the Parties has not begun its supervisory work. Treaty form gives States a public way to affirm shared commitments to human rights, democracy, and the rule of law in AI governance, while broad obligations and a flexible implementation model preserve wide domestic room for regulatory choice. Because AI carries exceptional public salience, and because the Convention is the first binding instrument in a field long governed by soft law, signing it confers a visible benefit on States at little regulatory cost, a benefit States visibly spend, citing the fact of signature as evidence of alignment with rights-protective AI governance even as ratification remains distant. Signature is therefore doing more expressive work than regulatory work. The essay develops this claim through the Convention’s design, the politics of delayed entry into force, and article 18 of the Vienna Convention on the Law of Treaties, which gives signature legal significance before full operation through the interim duty to respect the treaty’s object and purpose. The space between signature and the Conference of the Parties’ supervision reveals a separation among three elements in international AI governance, namely, legal form, acceptance of shared principles, and submission to international supervision.
Recommended Citation: Mujib Jimoh, The First AI Treaty as Expressive Hard Law: Signature, Supervision, and the Space Between, 50 Fordham Int'l L.J. Online 1 (2026).