Hallucinated AI intelligence nearly launched a US strike on a Chinese ship

Military aircraft were already in the air this spring, preparing for an armed operation against a Chinese vessel, when U.S. officials discovered the intelligence behind the mission had been hallucinated by an AI chatbot, CNN reported Friday. The operation was called off at the last minute, according to the report, avoiding a potential clash with China.

The false intelligence traced back to a single query. An analyst at Special Operations Command asked an AI chatbot to combine open source information with classified signals intelligence, according to CNN. The chatbot misread the ship’s cargo manifest and produced inaccurate findings. The analyst then used the same tool to turn those findings into a polished, official-looking summary, which spread through command channels as fact. The report claimed the vessel was carrying components for a nuclear weapons program, and it circulated during the war with Iran this spring.

The near-miss points to a problem military officials and outside experts have been warning about: as commanders and analysts lean more on AI tools, mistakes those systems produce can move up the chain of command before anyone checks them. The Pentagon has pushed to fold AI deeper into military operations, describing the technology as a way to speed up the “kill chain,” the sequence of steps from spotting a target to striking it. The same speed that makes AI appealing for that purpose can also let errors slip through without enough human review.

A call for safeguards, not retreat

Jake Steckler, a research scholar at GovAI and a veteran U.S. Army officer, said service members need to understand the uncertainty built into LLMs, the AI models behind chatbots. That understanding matters most, he told TechCrunch in a written response, “for any decisions that could lead to use of force, like targeting, intelligence analysis, or operational planning. There are life and death consequences for those decisions.”

Steckler said the episode should push the military toward stronger safeguards rather than away from AI altogether. “These tools can be useful in the right contexts and with the right safeguards in place,” he said. “But prioritizing adoption speed over all else will likely lead to incidents that only make service members lose trust in these systems, which ultimately is only going to slow adoption.”

CNN’s report, as relayed by TechCrunch, does not name the analyst, identify the chatbot used, or explain exactly how officials caught the error before the operation proceeded. Watch for whether the Pentagon responds with new verification steps for AI-assisted intelligence before it reaches commanders, and whether military oversight bodies seek a fuller account of how close the operation came to launching.