US and China experts propose shared limits on AI and nuclear weapons
Experts from the United States and China have proposed common rules to prevent artificial intelligence systems from independently making decisions about nuclear weapons. Melanie Sisson of the Brookings Institution and Tianjiao Jiang of Fudan University presented the recommendations before a planned September 24 meeting between Donald Trump and Xi Jinping.
The proposals identify several restrictions. AI systems should not be permitted to launch nuclear weapons autonomously or attack nuclear command systems. Human officials should retain exclusive authority over AI-assisted cyberattacks against strategic infrastructure. The two countries are also urged to establish a common definition of “human control.”
The recommendations build on an agreement reached by Biden and Xi in November 2024. Jiang additionally suggested creating a hotline for AI-related incidents. An automated defensive system could misinterpret suspicious activity and respond in a way that looks like an attack, while direct communication could help clarify an accidental event. The National Security Commission on Artificial Intelligence made a similar call in 2021, but binding arrangements have not been established. Carla Freeman of Johns Hopkins University questioned whether a hotline would work, citing China’s failure to answer a US call during the 2023 spy balloon crisis.
Written by Sakhanda Wire AI from reporting by The Decoder. Facts, quotes and full context are in the original article below.