The Trump administration has warned that Chinese AI companies could face U.S. sanctions if investigations confirm they developed AI models using intellectual property stolen from American firms. U.S. Treasury Secretary Scott Bessent said the government is examining whether some Chinese open-weight AI models were built through unauthorized use of U.S. technology, calling such practices "unacceptable" if proven.
The concerns center on a technique known as AI distillation, in which a smaller model learns from the outputs of a more advanced model. While distillation is a legitimate AI training method when properly licensed, U.S. officials and several AI companies allege that some Chinese developers may have used proprietary American models without authorization. Companies including Anthropic have previously accused Chinese AI labs of attempting to extract capabilities from their models through large-scale automated access.
If evidence of intellectual property theft is established, the administration could impose sanctions on the companies involved, potentially restricting their access to U.S. technology, financial systems, or international business opportunities. Officials are also considering additional disclosure requirements for U.S. organizations that deploy or integrate AI models found to have been developed using misappropriated intellectual property.
The warning marks another escalation in the growing AI rivalry between the United States and China, where competition increasingly extends beyond semiconductors to AI models and intellectual property. While U.S. officials argue that stronger enforcement is needed to protect innovation and national security, the move is expected to intensify debate over open-weight AI, global AI governance, and the balance between fostering competition and safeguarding proprietary technology.
