September 2026 international joint statement on frontier AI control

Multilateral effort, launched by a September 2026 joint statement from leaders of 20 countries and the European Commission, to place frontier AI models under coordinated human oversight through mandatory safety testing, common government standards, and a proposed international institution.