SOCOM AI Chatbot Nearly Triggered US Boarding of Chinese Ship

During the spring 2026 war with Iran, an intelligence report told US commanders that a Chinese ship in the Middle East was carrying components of a nuclear weapons program. Armed troops were preparing to board it. Military planes were already in the air. Just before the intercept, officials realized the report had been produced with help from an AI chatbot, and the chatbot had gotten the cargo wrong. One source told CNN the finding was "entirely false." Another said the episode "almost started a war."

The analyst worked at US Special Operations Command Pacific, and had asked a chatbot to fuse open-source intelligence with secret signals intelligence about the ship's manifest. CNN reports that "it was not clear whether the chatbot was a commercially available one or a US government product."

Defense Secretary Pete Hegseth has been publicly pushing hard on military AI. "We will unleash experimentation, eliminate bureaucratic barriers, focus our investments and demonstrate the execution approach needed to ensure we lead in military AI," he said in remarks CNN cites. A former senior US official was drier about what analysts actually get their hands on: "The internal tools are mostly just copies of the commercial stuff wearing lipstick."

The wider worry in the piece is targeting. A military source told CNN, "AI in targeting is definitely something that is ramping up and there is no real guidance for how having a human in the loop will prevent civilian casualties or fratricide." Another line lands harder: "AI allows you to get to a bad idea faster."

The near-incident surfaces the same week Brookings and Tsinghua scholars publicly called for nuclear-style AI safeguards and a US-China military hotline.