US Safety Tests to Exclude Chinese Open-Weight AI Models

Upcoming US safety testing regulations for artificial intelligence models will reportedly exclude open-weight models developed in China. This exemption directly impacts prominent Chinese open-weight models such as DeepSeek and Qwen. This policy decision could significantly shape the global open-source AI landscape and influence geopolitical competition between the US and China. By bypassing stringent US regulatory hurdles, Chinese open-weight models may see accelerated adoption and integration worldwide. While US-developed proprietary models face rigorous safety evaluations, Chinese open-weight models will not be subjected to the same US-mandated testing. This highlights a regulatory challenge in governing models whose weights are already publicly downloadable and distributed globally.