Chinese military reportedly uses American AI models to train its defense systems - tools from OpenAI and Anthropic reportedly among those affected

Chinese military researchers have reportedly used outputs from American AI models by OpenAI and Anthropic to train smaller, locally runnable defense systems through model distillation, bypassing US export controls on advanced chips. This method allows China to inherit complex reasoning capabilities without needing the restricted hardware, raising strategic concerns amid ongoing AI competition.