Can Cold War-era nuclear playbooks keep autonomous weapons from starting an unintended global conflict? As artificial intelligence moves from battlefield support to direct command operations, top security experts from the US and China warn that a single machine glitch could spark a catastrophic military escalation in minutes. Ahead of an upcoming summit between Donald Trump and Xi Jinping, researchers from the Brookings Institution and Fudan University are urging both superpowers to enact nuclear-style safeguards and strict human controls over military AI before automated systems make decisions no one can take back.
Cold War playbooks for modern machine speeds
The proposal was authored by Brookings senior fellow Melanie Sisson and Fudan associate professor Tianjiao Jiang. It highlights how quickly the US-China tech rivalry is shifting from chip tariffs to classic arms control. Because autonomous systems move faster than human reaction times, an automated response to a false alarm could easily look like a deliberate act of war.
To prevent miscalculations, the researchers recommend drawing strict red lines that forbid AI from independently launching nuclear weapons or targeting a rival nation’s nuclear command infrastructure. They want a shared, formal definition of meaningful human control. This should ensure both governments maintain actual human authority over consequential cyber operations rather than relying on automated decision-making.
The challenge of getting both sides to answer the phone
A central piece of the proposal involves creating a dedicated US-China military hotline specifically for AI incidents. If an autonomous system malfunctions, runs rogue, or experiences an unauthorized breach, a direct communication channel would allow military officials to clarify the situation before either side orders a retaliatory strike.
However, experts like Carla Freeman from Johns Hopkins University point out that hotlines only work if someone picks up. During previous geopolitical friction, such as the 2023 balloon incident, Chinese officials hesitated to respond until senior leadership finalized a stance—a bureaucratic process that can take days rather than the minutes required during an AI crisis.
Philosophical divides ahead of official talks
Neither government has officially adopted the proposals. Beijing handles AI under its Foreign Ministry’s arms-control unit, though. Meanwhile, Washington spreads AI security across the Pentagon, State Department, and National Security Council.
With official AI talks on the horizon, both nations remain cautious about signing agreements that might slow down their own technological momentum. Still, as AI shifts from secondary support to primary command systems, keeping final authority strictly in human hands may be the one area where both global powers share a common interest.
The post Could Autonomous AI Trigger an Accidental War? US and Chinese Experts Demand Red Lines appeared first on Android Headlines.
​Â
