A group of experts from the United States and China has proposed the formulation of common rules to exclude AI control from nuclear weapons systems and maintain operations under human oversight. This initiative aims to curb the unpredictable risks that rapidly developing AI technologies pose to nuclear disarmament and global security.
Rather than a specific product announcement, this initiative constitutes a policy recommendation toward establishing a consensus. It emphasizes the necessity of international norms that prohibit AI from making autonomous attack decisions and exercising control in nuclear weapon operations. The AI control of nuclear weapons remains a critical concern, as malfunctions or misinterpretations of data could inadvertently trigger an unintended nuclear conflict.
Through dialogue, experts from both nations are working to build trust by discussing guardrails to prevent runaway AI from both technical and political perspectives. Moving forward, it is expected that this recommendation will serve as a foundation for deeper international discussions, ensuring transparency in the military applications of AI and establishing legal frameworks that mandate human involvement (Human-in-the-loop) in nuclear weapon systems.