The US military came dangerously close to boarding a Chinese cargo ship in the Middle East this spring after an AI chatbot generated a fabricated intelligence report claiming the vessel was carrying nuclear-weapons components, CNN reported Thursday, citing multiple people familiar with the matter.
According to CNN's sources, a special-operations analyst queried an AI chatbot about a classified intelligence report concerning the Chinese ship. Whether the chatbot was a commercially available product or a US government system was not specified. The AI combined publicly available information with classified intelligence from government systems and arrived at its erroneous conclusion about the ship's cargo. The analyst then incorporated the AI's output into a standardized intelligence briefing -- the kind that military personnel trust and act on -- and circulated it.
Based on this report, the military mobilized. Plans were drawn up to intercept and board the ship. Armed military personnel were briefed and prepared. Military aircraft were already in the air when other personnel scrutinized the report more closely and identified the AI-generated portion. It became clear that the supposed cargo had been fabricated, and the operation was called off.
"The report was completely false" and "almost started a war," one source told CNN, noting that any military operation against a Chinese vessel could have escalated into a much broader conflict.
The incident highlights a growing tension in modern military operations: the US military is increasingly relying on AI to assist in selecting targets for strikes, but AI systems are prone to hallucinations -- generating plausible-sounding but entirely fabricated information. When that fabricated intelligence enters standardized military briefing formats and is acted upon at speed, the consequences can be catastrophic.
CNN did not report the exact date of the incident, stating only that it occurred in the spring of 2026.
