AI-fabricated intelligence nearly triggered a US military strike on China
This spring, US military aircraft were already airborne for an operation against a Chinese vessel when officials discovered the underlying intelligence had been invented by an AI chatbot, and the mission was called off at the last moment. The incident has fueled concerns about the military's growing reliance on AI for high-stakes decisions.
US military officials aborted an armed operation against a Chinese vessel at the last minute this spring after discovering that the intelligence behind it had been fabricated by an AI chatbot, CNN reported Friday, citing sources familiar with the matter. Aircraft involved in the mission were already airborne when the error was caught, narrowly avoiding a potential escalation with China.
The intelligence report, which circulated during the war with Iran, claimed the vessel was carrying components for a nuclear weapons program. That claim turned out to be false. It originated with an analyst at Special Operations Command who had used an AI chatbot to combine open-source data with classified signals intelligence. The chatbot misread the ship's cargo manifest. The same analyst then used the tool again to turn the flawed conclusion into a polished, official-looking report, which was then distributed through command channels.
Speed versus reliability
The near-miss comes as the US military pushes to adopt AI more broadly in order to speed up decision-making and keep pace with China. The Pentagon has said AI offers a major advantage by accelerating the "kill chain," allowing commanders to act within tighter timeframes. But that same speed, critics warn, can let AI-generated errors advance through the chain of command before anyone questions them.
Jake Steckler, a research scholar at GovAI and a US Army veteran officer, said service members need to understand the inherent uncertainty of large language models, particularly for decisions tied to the use of force, such as targeting, intelligence analysis, or operational planning, where mistakes carry life-and-death consequences. Still, Steckler said the episode should prompt stronger safeguards rather than a retreat from AI adoption. He warned that prioritizing speed of deployment above all else risks incidents that erode troops' trust in these tools, which would ultimately slow their adoption rather than accelerate it.
