According to Reuters, reported via The Next Web and Quartz, Chinese officials are weighing limits on who outside the country can use the nation's strongest AI models. The Ministry of Commerce has reportedly held talks with Alibaba, ByteDance and startup Z.ai, and any curbs could extend to open-weight models — the freely downloadable systems, including Alibaba's Qwen and Z.ai's GLM, that made Chinese AI popular abroad. Officials have also floated treating theft of proprietary AI as a national-security crime. The reports stress no decision has been made and any rules might apply only to future models. It would mirror the U.S. export controls that recently restricted Anthropic's frontier systems — now in reverse.