Washington, September 22, 2026, SPA — In a joint statement, leaders of more than 20 countries called for binding safety measures for advanced artificial intelligence (AI) systems. They warned that the systems could develop faster than they can be regulated and controlled.
The statement cited serious risks to safety and security that could arise from misuse of these systems, as well as recent cases in which AI programs bypassed testing safeguards and gained unauthorized access to real-world systems. It stressed that these systems must remain under human direction, oversight, and control, and that oversight measures must not widen the gap between countries in access to AI’s benefits.
The statement came on the eve of UN General Assembly meetings in New York, where AI’s growing role and its impact on jobs and cybersecurity are expected to draw wide attention.
…