Gebru, Bender: AI chatbot error nearly triggered US-China war

In spring 2026, during the war with Iran, a special operations command analyst asked an AI chatbot to help assess intelligence on a Chinese ship in the Middle East. The chatbot produced a mishmash of open-source information and secret government intelligence and concluded the vessel was transporting components of a nuclear weapons program. Military planes were in the air and armed members were preparing to board the ship before officials discovered the report had been generated with the help of AI and the material had been misidentified.

That incident, first reported by CNN on September 18, is the hook of a Guardian op-ed by Timnit Gebru, executive director of DAIR, and Emily M. Bender, a linguistics professor at the University of Washington, published October 1. One CNN source said the AI "almost started a war" and called the "hallucination" part of a trend inside the US military rather than an isolated case. CNN said it could not determine what the ship was actually carrying, or whether the analyst's tool was a commercial product or a government-designed system.

"The U.S. almost started World War III with China, relying on a faulty, fabricated information from a chatbot," Gebru said in a related Democracy Now interview. "They are fallible, they're not superintelligent."

The authors contrast the thin coverage of that near-miss with the attention paid to a resigning Anthropic researcher cited in the piece who warned about companies "racing straight towards self-improving superintelligence and gambling with our lives." Their argument is that the real danger is error-prone systems already being pushed into high-stakes decisions. Gebru's prescription: "We need to hold corporations liable for unleashing these kinds of error-prone models into the world."