Establishing an independent multilateral governance body to mitigate geopolitical, socioeconomic, and existential risks from powerful artificial intelligence should be an urgent global priority.

Field-founding scientists, experts, and researchers at frontier labs agree that the potential of AI to bring immense benefits and catastrophic risks is underestimated. Rapid capability development may lead to the emergence of unexpected features — outside the scope of safety mechanisms and beyond human understanding — which could accelerate interconnected risks. Immediate action is essential to prevent plausible human extinction.