This spring, during the war with Iran, an intelligence report circulated across the US military claiming that a Chinese ship transiting West Asia was carrying components of a nuclear weapons programme. The report immediately set off alarm bells.

Four people familiar with the episode told CNN that the US military quickly prepared to intercept the vessel. Armed service members were preparing to board it, and military aircraft were already in the air. But just before the operation began, officials took a closer look at the intelligence and discovered that it had been generated with the help of an AI chatbot, which had misidentified the ship’s cargo. What actually went wrong?