Chinese Military Uses U.S. AI Models for Defense Systems

Chinese military researchers have been using outputs from advanced AI models developed by U.S. companies OpenAI and Anthropic to train domestic AI systems, according to a Reuters review of over 80 Chinese academic papers and patents. This technique, known as 'model distillation', allows for the development of smaller, specialized AI systems that can operate locally with fewer computational resources. The practice has raised concerns in Washington about intellectual property rights and export controls, while China denies any wrongdoing.