The Trump administration will look into whether Chinese artificial intelligence (AI) models have been stolen or distilled from American models, U.S. Treasury Secretary Scott Bessent said on Tuesday.
“This administration supports open-source models, but what we do not support is IP theft,” Bessent said on Fox Business on July 21. “If we see, especially, that overseas models are stealing from our great companies, we have the ability to sanction them because of this theft.”
“We are finding watermarks of our U.S. large language models on many of the Chinese models, and that’s unacceptable,” he added. “So, we’re going to be looking at that in the coming days or weeks.”
Bessent’s comments follow the recent release of Chinese open-source models, such as Moonshot AI’s Kimi K3 model and Alibaba’s Qwen 3.8 Max, which have gained popularity globally.
American AI executives are warning that these models could pose vast security risks and threaten the business models of top U.S. AI firms like OpenAI and Anthropic, according to a report from The Wall Street Journal.
Bessent explained that this kind of theft is called distillation, a training method in which a smaller model learns from the outputs of a larger model, to replicate the capabilities of leading U.S. AI systems.
“These models have to be held to the same standard that U.S. models are,” Bessent said.
“They have hundreds of millions of hits against the large language models here in the U.S. and they try to reconstruct their code and the thought process of the models, and that’s unacceptable,” Bessent said, adding, “You can’t use counterfeit goods.”
Bessent also raised the possibility of requiring companies that deploy Chinese AI models to disclose that information to customers, though he did not provide additional details on whether the administration is considering formal disclosure requirements.
The United States and China are planning to hold talks on AI in September, according to reporting from Reuters. Bessent will represent the United States during those discussions.