Canada, Kenya, the Netherlands, Australia, South Africa, Turkey, the Emirates... The September 21 declaration calls for independent tests before deployment. Washington, London and Beijing are not included.
On the eve of the opening of the United Nations General Assembly, on September 21, 2026, twenty states signed a joint declaration calling for international oversight of artificial intelligence. Among them: Canada, Kenya, the Netherlands, Australia, South Africa, Turkey and the United Arab Emirates. The United States, the United Kingdom and China are not among them.
The text consists of three demands. First, to subject AI systems to tests conducted by independent evaluators before they are made available to the public. Second, to define common safety standards at the international level. Finally, to create a global institution capable of setting standards, allowing them to be verified and convening States when technological thresholds are crossed.
The guiding principle is that AI must remain under human direction, supervision and control, and be developed in accordance with international law. The signatories justify the urgency by the speed of technical progress and by recent episodes where systems have circumvented the safety protocols that are supposed to govern them.
Read the full article on the original source
Go to the source →