Brandon Fletcher
CNN reports that the US nearly initiated military action against a Chinese vessel suspected of transporting nuclear weapon components.
A misleading intelligence report, partly generated by artificial intelligence, nearly incited an attack by the US military against a Chinese ship located in the Middle East, as detailed by CNN. This incident, described as having the potential to “almost start a war,” underscores the significant dangers associated with integrating AI into the critical operations of the US Department of Defense.
Engadget has reached out to the Department of Defense for comments regarding CNN‘s findings. We will provide updates on this article once we receive a response.
An analyst from the special operations command compiled the assessment by querying an AI chatbot about the manifest of the Chinese vessel. As reported by CNN, the AI system synthesized “open-source intelligence with classified signals intelligence” to assert that the ship was carrying “components of a nuclear weapons program.” This erroneous conclusion almost prompted military action to intercept the vessel before officials confirmed the information was generated by AI.
CNN was unable to ascertain the actual cargo of the ship or identify the specific AI tool used by the analyst, whether it was a commercial product or one developed for the US government. With the Department of Defense actively pursuing AI integration, several possibilities exist for the technology employed.
In January, Secretary of Defense Pete Hegseth issued a memo urging the DoD to adopt an “AI-first” approach and to experiment with models from leading American AI firms. This initiative faced challenges, particularly with Anthropic, which declined to allow the military to utilize its models for autonomous weapon development, leading to a temporary ban by the government. Nevertheless, the effort has fostered multiple AI partnerships. Notably, SpaceXAI reportedly agreed to provide the military access to Grok in February, while NVIDIA, Microsoft, and Amazon formed a partnership to deliver their AI technologies to the DoD in May. Such government contracts are highly lucrative, attracting corporate interest. the expansion of AI applications involves risks that differ from traditional enterprise software solutions.
Daniel Mercer
