False AI-generated intel nearly led US military to attack Chinese ship, CNN reports

CNN has reported that a false intelligence report partly generated by an AI chatbot nearly prompted the US military to attack a Chinese ship in the Middle East that was believed to be carrying nuclear weapon components. The incident 'almost started a war,' according to a source cited by CNN, underscoring the risks of integrating AI into sensitive Department of Defense work.

According to CNN, a special operations command analyst asked an AI chatbot about the ship's manifest. The chatbot combined open-source intelligence with secret signals intelligence and concluded the vessel contained 'components of a nuclear weapons program' – a finding that almost led the military to apprehend the ship until officials realized the assessment was AI-generated. CNN could not determine what the ship was actually carrying or which AI tool, commercial or government-designed, was used.

The incident comes as the DoD pushes to adopt AI more broadly. In January 2026, Defense Secretary Pete Hegseth issued a memo directing the department to become 'AI-first' and experiment with models from leading US AI companies. Anthropic initially refused to allow its models to be used for autonomous weapons development and was briefly banned by the government. Other deals have since emerged: SpaceXAI reportedly agreed in February 2026 to let the military use Grok, and in May 2026 Nvidia, Microsoft and Amazon partnered to offer AI technologies to the DoD.

The report notes that such government contracts are lucrative, but the near-miss illustrates that AI carries risks beyond those of conventional enterprise software.