The US Treasury Secretary Scott Bessent announced on Tuesday that the United States government will scrutinise Chinese open source AI models for potential intellectual property (IP) theft. This move signals a significant escalation in the ongoing technological competition between the US and China in the AI sector.
Bessent stated on Fox Business that while the administration supports open source AI models, it will not tolerate IP theft. He warned that if evidence emerges of Chinese AI companies stealing technology from US firms, sanctions could be imposed. This announcement comes amid rising concerns over Chinese AI models like Moonshot AI's Kimi K3, which are rapidly advancing and challenging the dominance of American AI companies such as OpenAI and Anthropic.
Growing Concerns Over IP Theft and AI Competition
The US has already implemented restrictions on China's access to advanced semiconductor chips and tightened export controls. Now, the government is considering targeting the AI models themselves. This approach aims to protect US firms’ investments and intellectual property in the face of increasingly sophisticated Chinese AI developments.
However, the issue of what constitutes IP theft in AI remains contentious. Techniques like model distillation, which involves creating smaller AI models based on larger ones, are widely used and debated. Microsoft CEO Satya Nadella recently criticised the assumption that distillation equates to theft, highlighting the complexity of AI training and innovation.
Industry Perspectives on China’s AI Progress
Some experts argue that China’s AI advancements are not solely due to copying or distillation. Clem Delangue, CEO of Hugging Face, noted that China has strong research teams adopting a more open and collaborative approach to AI development compared to the US. This suggests that China’s progress is driven by genuine innovation alongside strategic use of open source methods.
Meanwhile, US AI companies continue to face legal risks related to their own training practices. For example, Anthropic recently received judicial approval to compensate authors after settling claims over the unauthorised use of copyrighted books in AI training.
Looking Ahead
The US government’s potential sanctions on Chinese AI models reflect broader efforts to maintain leadership in AI technology. These measures underscore the intersection of policy, innovation, and international competition shaping the future of AI development.
For businesses and developers navigating this evolving landscape, practical AI adoption and compliance with IP regulations will be crucial. Solutions like those offered by can help organisations scale AI responsibly while managing legal and ethical considerations.
Scope and Implementation Disclaimer: This article provides an overview of policy developments and industry perspectives related to AI and intellectual property. It does not offer legal advice or specific guidance for compliance. Readers should consult appropriate experts when addressing IP or regulatory issues in AI projects.
For JASON AI services, AI News, and practical AI installation guidance, visit https://jasonjuul.com.