U.S. Treasury warns sanctions possible against Chinese open-source AI models over alleged IP theft

On Tuesday, U.S. Treasury Secretary Scott Bessent said the government would examine open-source AI models developed in China for signs of intellectual property theft, and that sanctions could follow if violations are established. Bessent made the remarks during an interview on Fox Business, stating that while the administration supports open-source models, it does not condone theft of proprietary capabilities. "If we see, especially, that overseas models are stealing from our great companies, we have the ability to sanction them because of this theft," he said.

The warning comes as Chinese open-source models such as Moonshot AI’s Kimi K3 have rapidly increased in capability and adoption, raising concerns among U.S. firms like OpenAI and Anthropic about competitive pressure and capital formation for frontier model development.

Earlier on Monday, Axios reported that the Trump administration was considering a broader ban on Chinese open-source models, though the claim has been disputed by other sources.

Bessent’s comments follow a White House pledge in April to work with AI companies to combat technology theft. The administration has already restricted China’s access to advanced semiconductors and tightened export controls, and now appears to be considering measures that would directly target AI models—a potential escalation in the U.S.-China technological competition.

The move raises legal and technical questions about what constitutes model theft, particularly in the context of model distillation, a technique used to transfer capabilities from larger to smaller models. Microsoft CEO Satya Nadella recently criticized the assumption that distillation inherently constitutes theft, noting that many labs rely on fair use rights to train on public data but then impose restrictive terms on downstream distillation.

Industry leaders also argue that distillation is not the main driver of China’s progress in AI. Hugging Face CEO Clem Delangue said that distillation is a minor factor in building high-performing models and is widely practiced, including in the U.S. He attributed China’s advances to strong research teams and a more open, collaborative approach to AI development.