A chatbot hallucination almost triggered a US military attack on a Chinese ship

An AI hallucination almost sparked an international military crisis.

The US military was preparing air support to intercept and board a Chinese cargo ship in the Middle East. The operation was cancelled after officials realized the intelligence report behind it was entirely false. According to CNN, a US Special Operations Command analyst used a chatbot to analyze the ship's manifest by combining open-source and classified signals intelligence. The AI inaccurately reported that the vessel carried nuclear arms program components, leading one source to state the mistake "almost started a war."

Why it matters: AI hallucinations are no longer just messing up legal briefs or student papers—they are reaching high-stakes defense decisions. Despite these risks, the Department of Defense rolled out an "AI acceleration strategy" in January to feed more military data into AI tools across every service branch.

Double-checking your AI's work is good practice; doing it before sending armed forces to board a foreign ship is essential.