US President Donald Trump has called the competition over artificial intelligence the most critical contest in centuries. Yet his own administration is now carving out a surprising exemption in its new AI safety framework, proposing to exempt open-weight AI models from mandatory US safety reviews.

The move could inadvertently benefit China's rapidly advancing open-weight AI models, including DeepSeek and Moonshot AI's Kimi K3. CGTN's Liu Jiaxin speaks with industry experts to examine the potential implications.