Full Breakdown
AI Risks Prompt Nuclear-Style Safeguard Proposals
By Drooid · · How we work
Core Proposal Overview
U.S. and Chinese security experts participating in a Brookings-Tsinghua dialogue have issued a set of safeguards aimed at preventing autonomous artificial-intelligence systems from triggering nuclear or strategic-infrastructure crises. The recommendations call for explicit red lines that prohibit AI from independently initiating cyberattacks against nuclear command, control and communications (C3) networks or from deciding to use nuclear weapons. They also call for a shared definition of “meaningful human control” and the creation of a dedicated U.S.–China military hotline to report accidental or unauthorised AI incidents before they are interpreted as hostile actions.
Background & Context
The dialogue, convened since 2019, builds on a principle endorsed by former U.S. President Joe Biden and Chinese President Xi Jinping in November 2024 that humans must retain authority over nuclear-weapon use. Melanie Sisson, senior fellow at the Brookings Institution, and Tianjiao Jiang, associate professor at Fudan University, authored the latest set of proposals. China has formally placed AI within the Foreign Ministry’s Department of Arms Control—a portfolio that also handles nuclear weapons, missiles and non-proliferation—while the United States lacks a single arms-control home for AI, with policy dispersed across the National Security Council, State Department, Pentagon and other agencies.
Official Statements & Responses
Shen Jian, China’s ambassador for disarmament affairs, told a United Nations meeting in Geneva in June that military AI could affect strategic stability and raise the risk of miscalculation, emphasizing that weapons must remain under human control. Both governments have not publicly endorsed the Brookings-Tsinghua proposals, and each remains cautious about measures that might impede its own technological development.
What’s Next
The proposals were released ahead of an expected Sept 24 meeting in Washington between U.S. President Donald Trump and Chinese President Xi Jinping, where government-level AI talks are slated to occur. The outcome of that dialogue will indicate whether the suggested nuclear-style safeguards gain official traction.