China didn’t need to steal America’s best AI. It just had to talk to it. According to a Reuters investigation published Thursday, Chinese military researchers have been feeding answers from leading US AI models, including OpenAI’s and Anthropic’s — into their own systems, using them to sharpen China’s defence capabilities. The findings, drawn from more than 80 academic papers and patents, expose a workaround to Washington’s years-long effort to keep cutting-edge technology out of Beijing’s reach.
