Treasury Secretary Warns China Could Face Sanctions Over AI Model 'Theft'

U.S. Treasury Secretary Scott Bessent said Tuesday the Trump administration is investigating whether Chinese artificial intelligence models were built by copying American ones, calling the practice "IP theft" and warning that sanctions could follow.

Chinese open-weight AI models are increasingly competitive with top U.S. offerings from companies like OpenAI and Anthropic. Moonshot AI's new Kimi K3 model, released this month, beats American rivals on several industry benchmarks. Bessent told Fox Business that if overseas models are "stealing from our great companies, we have the ability to sanction them." He described the technique as "distillation," where a smaller model is trained using outputs from a larger, existing model. Anthropic recently alleged in a letter to the U.S. Senate that Chinese tech giant Alibaba carried out the largest known distillation attack against it.

The U.S. and China are scheduled to hold AI talks in September, with Bessent representing the U.S. "We are finding watermarks of our U.S. large language models on many of the Chinese models, and that's unacceptable," Bessent said, adding that his team will examine the issue "in the coming days or weeks."

Both Anthropic and OpenAI have faced their own accusations of intellectual property theft. In September, Anthropic agreed to pay $1.5 billion to settle a class action lawsuit from authors who said the company downloaded books from pirated databases. The New York Times sued OpenAI and Microsoft in 2023, alleging copyright infringement for using its articles to train AI models.

Source: CNBC