In a historic ruling, Chinese President Xi Jinping and U.S. President Joe Biden agreed that human decision-makers, not artificial intelligence, should continue to make judgments on nuclear weapons. After their meeting on Saturday, the White House released a statement emphasizing this accord. Given the possible hazards, both leaders underlined the need for caution in the military’s development of AI technologies.
Similar comments were expressed by the Chinese government, albeit it did not immediately offer any additional remark. Although the agreement is a big start toward tackling the relationship between AI and nuclear weapons, it’s unclear if it will result in official talks or tangible actions.
This accord represents a rare instance of collaboration between the two global powers, especially on contentious issues such as nuclear arms control and AI governance. The United States has long urged China to participate in nuclear arms talks, with minimal success. Although semi-official discussions briefly resumed in November, formal negotiations remain stalled. U.S. officials have expressed concern over China’s rapidly growing nuclear arsenal, which the Pentagon estimates will exceed 1,000 warheads by 2030. This growth contrasts with the current operational warhead counts of 1,770 and 1,710 for the U.S. and Russia, respectively.
In recent years, China has improved its nuclear triad, tested hypersonic glide vehicles, and introduced next-generation ballistic missile submarines as part of its nuclear modernization efforts. Beijing maintains a “no first use” policy and supports modest deterrence in spite of these developments. Chinese scholars have reaffirmed their nuclear stance and rejected Western concerns as irrational.
The importance of these conversations is highlighted by the agreement on human control of nuclear decisions. This agreement may open the door for greater cooperative efforts on international security and responsible AI use, even though the wider ramifications are still unknown.
