Multiple reports say Chinese military-linked researchers have used outputs from leading U.S. artificial intelligence models to train domestic defense-related systems. The accounts describe “model distillation,” a technique in which an AI system is trained to replicate or learn from the responses produced by another model. According to the reports, researchers have drawn on U.S. AI models associated with OpenAI and Anthropic to develop or refine Chinese AI capabilities intended for military applications.
Reuters-referenced reporting cited by one outlet says more than 80 Chinese papers and patents are reviewed showing use of distillation methods by institutions with military ties. The described target areas include surveillance, cyber operations, and battlefield applications. The reports frame the activity in the context of intensifying technology and strategic competition between the United States and China.
While the sources align on the general approach and the use of U.S.-developed model outputs, they vary in emphasis and in the breadth of the documented materials. No source provides conclusive public details on the specific operational systems fielded by the Chinese military.
