Bessent Says U.S. May Sanction China Over AI Model Distillation Theft

U.S. Treasury Secretary Scott Bessent stated on Tuesday that the Trump administration may impose sanctions on China over alleged artificial intelligence model theft through a process called distillation. Bessent told Fox Business the government does not support intellectual property theft, specifically citing concerns that Chinese AI models may have been built using outputs from American models. The statement comes as Chinese open-weight models like Moonshot AI's Kimi K3, released earlier this month, gain competitive ground against leading U.S. offerings from companies including OpenAI and Anthropic, raising questions among tech executives and officials about America's position in the AI development race.

Anthropic Alleges Alibaba Conducted Distillation Attack

Bessent explained that distillation is an AI training method where a smaller, less capable model is built using outputs from an existing, stronger model. "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 on Tuesday. Anthropic sent a letter to the U.S. Senate Committee on Banking, Housing, and Urban Affairs last month alleging that Chinese tech company Alibaba had carried out "the largest known distillation attack" against it to date. "We are finding watermarks of our U.S. large language models on many of the Chinese models, and that's unacceptable," Bessent said Tuesday. "We're going to be looking at that in the coming days or weeks."

U.S. and China Plan AI Talks in September

The U.S. and China are planning to hold talks around AI in September, according to a Reuters report on Tuesday. Bessent will represent the U.S. during the discussions, the report said.

FAQ

What did Treasury Secretary Bessent say about Chinese AI models on Tuesday? Bessent stated on Tuesday that the Trump administration may sanction China over alleged AI model theft through distillation, citing evidence of U.S. large language model watermarks found on Chinese models.

What is AI model distillation according to Bessent? Bessent explained that distillation is an AI training method where a smaller, less capable model is built using outputs from an existing, stronger model. Anthropic alleged in a letter last month that Alibaba conducted "the largest known distillation attack" against it.

When are U.S.-China AI talks scheduled? According to a Reuters report on Tuesday, the U.S. and China are planning to hold talks around AI in September, with Bessent representing the U.S. during these discussions.