"Almost Started a War": US Fighter Jets Were Already in the Air — Then the AI Error Was Revealed

A faulty intelligence report produced with the help of artificial intelligence nearly led the US military to attack a Chinese ship in the Middle East during the war against Iran. The incident was revealed by CNN and raises questions about reliance on AI in military intelligence.

A faulty intelligence report circulated last spring within the US military, amid the war against Iran, nearly led to a military confrontation between the US and China. According to a CNN exposé, the report determined that a Chinese ship sailing in the Middle East was carrying components of a nuclear weapons program. The US military quickly went into action and began planning an interception of the vessel, with fighter jets already in the air. A last-minute check revealed that the report was completely false, originating from an error by an artificial intelligence system. An analyst fed a chatbot data from the Pacific Special Operations Command, and the bot combined open-source intelligence with top-secret signals intelligence, reaching a wrong conclusion. The analyst then used AI again to package the findings into an official report, a format that receives automatic trust. A former senior US administration official sharply criticized the internal tools, describing them as copies of commercial products. The incident raises serious questions about the growing use of AI in intelligence and on the battlefield, and about the phenomenon of AI hallucinations, which is not an isolated event in the intelligence community. The pressure to produce fast intelligence opens the door to errors, especially among young analysts who tend to trust technological tools without careful scrutiny.