Gordon Chang warns of a troubling development in global security: China’s potential use of artificial intelligence (AI) to control nuclear weapons. This shift not only escalates tensions but also poses significant risks to global stability.
Rejection of AI Limitations
In recent discussions in Geneva, China rejected U.S. proposals to limit the role of AI in nuclear launch decisions, highlighting a growing rift between the two powers. President Biden had previously raised this issue with Xi Jinping, emphasizing that AI was an area for potential collaboration. However, as Tarun Chhabra, the director of technology at the National Security Council, stated, “Our position has been publicly clear for a very long time: We don’t think that autonomous systems should be getting near any decision to launch a nuclear weapon.” This rejection underscores Beijing’s commitment to advancing its military capabilities without external constraints.
The Perils of AI in Nuclear Systems
The integration of AI into nuclear arsenals introduces unprecedented risks. Historical incidents, like the one involving Soviet officer Stanislav Petrov in 1983, illustrate the critical role of human intuition in averting nuclear catastrophe. When alarms indicated a U.S. missile launch, Petrov relied on his instincts rather than protocol, ultimately recognizing the false alarm. Such human judgment could be absent in AI systems, potentially leading to automatic retaliatory strikes based on erroneous data.
As Brandon Weichert noted, “AI-controlled nukes would be far more devastating than your ordinary ones. The Chinese are behaving foolishly, though, because the chance for unintended consequences is high.” This highlights the unpredictable nature of AI in life-or-death situations.
A New Arms Race
China’s refusal to engage in discussions about limiting AI in nuclear contexts aligns with its broader strategy to achieve nuclear superiority. Richard Fisher, Jr. from the International Assessment and Strategy Center, pointed out that “the Chinese Communist Party has advocated nuclear arms control only for other nuclear powers.” This stance suggests that China seeks to enhance its strategic position through technological advancements.
The potential for AI to enhance initial nuclear strikes is particularly alarming. AI systems could identify optimal times and targets for attacks, increasing the likelihood of preemptive or retaliatory nuclear engagements. This development signals a shift from traditional nuclear deterrence to a more aggressive posture.
The risk of an AI controlled nuclear weapon cannot be calculate there are too many variables. I suspect that China is more interested in keeping its options open than actually doing this – and they have no problem busting our balls with it.