The administration considers banning Chinese open-weight AI models to protect American trade secrets and national security.
The administration is currently evaluating whether to restrict or ban Chinese open-weight AI models, such as Moonshot AI’s Kimi K3, to safeguard American intellectual property and national security. While Chinese labs have successfully used 'distillation' to bootstrap their models, the administration is distinguishing between legitimate technical training and fraudulent 'distillation attacks' where Chinese firms systematically extract U.S. trade secrets. To address these risks, the administration is considering sanctions for labs caught in industrial-scale fraud, such as Moonshot AI and Alibaba, while maintaining support for open-source software. Officials like Treasury Secretary Scott Bessent have warned of potential sanctions for intellectual property theft, while the Federal Communications Commission recently banned Chinese humanoid robots due to alleged national security risks. The administration aims to balance these actions by fostering American open-weight alternatives to ensure the U.S. does not lose its technological edge to China's state-subsidized diffusion. By punishing fraud rather than the underlying technique, the administration seeks to protect the interests of American frontier labs like OpenAI and Anthropic.