US and Chinese experts propose limits for military AI — Jerusalem Post
Experts from the United States and China have urged Washington and Beijing to prepare for incidents involving autonomous artificial intelligence systems. Such systems could interfere with nuclear command networks or conduct military cyber operations, leaving governments only minutes to determine whether these actions were accidental, unauthorized, or ordered by the other side, The Jerusalem Post reports.
The recommendations were prepared by Melanie Sisson, a senior fellow at the Brookings Institution, and Tianjiao Jiang, an associate professor at Fudan University. They participate in the US-China Dialogue on AI and National Security, conducted since 2019 by the Brookings Institution and the Center for International Security and Strategy at Tsinghua University.
Human control over cyberattacks
Sisson proposed reserving for humans the exclusive right to initiate AI-enabled cyberattacks against another country's nuclear command, control and communications systems or its strategically important infrastructure. According to the experts' assessment, an autonomous system could malfunction, exceed its instructions, or be compromised, while the side subjected to an attack might not immediately determine its cause.
More current news is available on the UA.News Telegram channel Telegram.
Jiang proposed establishing clear limits that would prevent AI from independently making decisions on the use of nuclear weapons or autonomously attacking nuclear command systems. He also advocated protecting critical infrastructure, including the energy sector, financial sector and healthcare, and jointly defining the concept of “meaningful human control.”
Hotline for AI incidents
Jiang also proposed creating a separate military hotline between the United States and China for incidents involving AI. He warned that an automated exchange of cyber actions could develop too quickly for human intervention: one country's defense system could respond to suspicious activity, while the other side could perceive that response as an attack and retaliate before the circumstances are clarified.
Neither government has publicly endorsed the proposals of the Brookings Institution and Tsinghua University. At the same time, China's ambassador for disarmament, Shen Jian, said in June at a UN meeting in Geneva that military AI could affect strategic stability and increase the risks of miscalculation and conflict escalation.
