US Could Sanction China Over AI Model 'Theft', Says Treasury Secretary

U.S. Treasury Secretary Scott Bessent indicated on Tuesday that the Trump administration is prepared to examine claims of intellectual property theft concerning artificial intelligence models originating from China. The potential for sanctions was raised if Chinese AI models are found to have been distilled from U.S. proprietary technology.

This development comes as Chinese open-weight AI models gain traction, challenging established offerings from American tech giants like OpenAI and Anthropic. Moonshot AI, a Chinese startup, recently launched its Kimi K3 model, which has reportedly surpassed U.S. models in certain industry benchmarks, raising concerns about the ongoing AI race.

"If we see, especially that overseas models are stealing from our great companies, we have the ability to sanction them because of this theft," Bessent told Fox Business. He added that the U.S. has found identifying marks, or watermarks, from American large language models within several Chinese models, a practice he deemed unacceptable.

Preparations are underway for U.S.-China talks on AI scheduled for September, with Bessent set to represent the U.S. The investigation into alleged AI model distillation follows similar concerns previously raised by companies like Anthropic, which reported an alleged "distillation attack" by Chinese firm Alibaba. Notably, both OpenAI and Anthropic have themselves faced accusations of copyright infringement related to the data used for training their AI models.