AI Error Nearly Triggered War Between US and China, Report Says
A US military analyst's use of artificial intelligence to assess a Chinese vessel's cargo led to plans for interception, which were halted when the AI's findings were proven false.
The U.S. military narrowly avoided a potential conflict with China earlier this year when an artificial intelligence system incorrectly identified the contents of a Chinese vessel, leading to U.S. military planning for an interception.
According to sources cited by CNN, a U.S. special operations analyst used an AI chatbot to examine the manifest of a Chinese ship traversing Middle Eastern waters. The AI erroneously concluded that the vessel was carrying components for a nuclear weapons program. This determination was then used to create an intelligence report that circulated within the military.
Following the report, military planners rapidly developed plans to intercept the ship. U.S. aircraft were mobilized, and soldiers prepared for a boarding mission. However, before the operation could be executed, officials reviewed the intelligence assessment and discovered it had been generated with AI assistance. They subsequently identified the AI's misidentification of the ship's cargo.
One source described the AI report as "entirely false" and stated that the incident "almost started a war" with China, as any military action against the vessel could have escalated into a conflict between the two global powers. The revelations come shortly before Chinese President Xi Jinping is scheduled to visit Washington for discussions with President Donald Trump on topics including AI, trade, and critical minerals.
The Trump administration has promoted the use of AI across federal agencies, with the Pentagon being a significant adopter of these technologies. AI has been employed for analyzing intelligence, identifying targets, streamlining logistics, and managing budgets. However, the incident highlights the risks associated with the rapid deployment of AI in critical intelligence functions, particularly when human oversight is insufficient.
It remains unclear whether the AI system used was an internal Pentagon tool or a commercial product. Sources suggest that internal military AI tools are often based on commercial technology. The Secretary of War has previously advocated for the swift integration of AI into the armed forces, an initiative that intensified earlier this year.
This event underscores concerns raised by AI companies like OpenAI and Anthropic regarding the potential for AI systems to generate inaccurate or misleading information, and the need for caution in their application, especially in high-stakes military operations.