AI error nearly started World War III, CNN reveals
An AI-generated intelligence report containing false information nearly pushed the U.S. military into an operation that could have triggered a global crisis.
A U.S. military intelligence report generated with artificial intelligence and riddled with errors nearly triggered an international crisis that could have escalated into a new world war, CNN revealed.
The incident occurred in the spring, when U.S. military officials launched an operation to intercept a Chinese ship in the Middle East under the suspicion that it was carrying materials used to produce nuclear weapons. According to multiple sources who spoke with CNN , military personnel prepared to board vessels while combat aircraft were already in the air.
However, before the operation moved forward, officials discovered that the threat did not exist. The information that had triggered the military response had been generated by artificial intelligence.
The report that sparked the alert was created with the help of a chatbot, but the tool incorrectly identified the cargo being transported by the ship. CNN was unable to confirm what materials the vessel was actually carrying or determine whether the system used was a commercially available chatbot or a version developed internally by the U.S. government .
A military source described the document as "completely false" and said the inaccurate information "almost started a war."
The incident highlights the risks of the U.S. government's growing reliance on artificial intelligence tools, a technology that can still generate false responses or fabricate information when it misinterprets data.
The Donald Trump administration has strongly pushed for the adoption of artificial intelligence within the defense apparatus. In July 2025, the government awarded Google a $200 million contract to develop AI solutions for the Department of Defense.
Months later, Defense Secretary Pete Hegseth introduced GenAi.Mil , an "American-made" artificial intelligence platform designed to help military personnel conduct research, create documents, and analyze images or videos at a faster pace.
Although it has not been confirmed whether that tool was involved in the false report about alleged nuclear materials, the case has raised concerns about the dangers of delegating sensitive decisions to systems that are still capable of making mistakes.
In January 2026 , Hegseth reinforced the Pentagon's strategy with a plan to accelerate the use of artificial intelligence across the armed forces and later claimed that the United States was using "every AI tool" available during the conflict with Iran.
