发生了什么
U.S. and Chinese security experts have proposed AI safeguards to prevent nuclear accidents and ensure human control over consequential cyberattacks. The proposals include red lines around nuclear systems, continued human control over cyberattacks, and a hotline for incidents involving autonomous AI.
The proposals were published last week ahead of planned government-level AI talks and an expected September 24 meeting in Washington between U.S. President Donald Trump and Chinese President Xi Jinping.
Melanie Sisson, a senior fellow at the Brookings Institution, and Tianjiao Jiang, an associate professor at Fudan University, authored the recommendations.
Their proposals reflect how the U.S.-China AI rivalry is expanding beyond competition over chips and advanced models. Increasingly, the discussion also involves issues traditionally linked to arms control.
China's State Security Minister Chen Yixin wrote this week in a government outlet that AI 'fundamentally transforms the form of military struggle'. He said AI had shifted from a supporting battlefield role to a key role, citing AI-directed strikes in the U.S.-Israeli war with Iran.
为什么这很重要
The proposals reflect growing concerns over increasingly powerful and autonomous AI systems. If implemented, they could help prevent accidents from being mistaken for attacks, limit machine-made military decisions, and establish communication protocols when technology behaves unexpectedly between nuclear-armed rivals.
The proposals build on a principle endorsed by then-U.S. President Joe Biden and Xi in November 2024 that humans should retain control over decisions to use nuclear weapons.
Lora Saalman, an associate senior fellow at the Stockholm International Peace Research Institute, wrote in February that this represented an important area of U.S.-China convergence.
At the same time, she noted signs that both countries were integrating AI more deeply into nuclear and nuclear-related systems.
Jiang also proposed a dedicated U.S.-China military hotline for AI incidents. He warned that automated cyber exchanges could escalate too quickly for humans to intervene.
互动机制：它实际上是如何运作的
以交互方式探索这一发展背后的基础技术。
crm_get_transaction(id='4092').How does the guide frame AI safety risk across current and advanced systems?
接下来看什么
The U.S. and China will hold government-level AI talks and a meeting between President Trump and President Xi Jinping on September 24. The proposals will be discussed in the context of the U.S.-China AI and national security dialogue.
The U.S. and China will hold government-level AI talks and a meeting between President Trump and President Xi Jinping on September 24.
The proposals will be discussed in the context of the U.S.-China AI and national security dialogue.
China has increasingly incorporated AI into its existing arms-control bureaucracy.
Washington has no equivalent single arms-control body responsible for AI.