USA Threatens Sanctions Against Chinese AI Models Over IP Theft

The United States has warned that Chinese artificial intelligence companies could face sanctions if investigations determine that their open-source AI models were developed using intellectual property stolen from U.S. firms. The warning, issued by U.S. Treasury Secretary Scott Bessent, marks the latest escalation in the AI rivalry between Washington and Beijing as concerns grow over the alleged use of model distillation and other techniques to replicate the capabilities of leading American AI systems.

Bessent said the U.S. government supports open-source AI development but would not tolerate the theft of American intellectual property. While he did not identify specific companies that could be sanctioned, his remarks follow repeated allegations by U.S. AI developers that several Chinese labs have trained advanced models using outputs from frontier American models without authorization.

US Signals Tougher Action on Alleged AI IP Theft

Speaking in an interview with Fox Business, Bessent said U.S. authorities would examine Chinese open-source AI models for evidence of intellectual property infringement.

According to the Treasury Secretary:

• The U.S. supports open-source AI development.
• Intellectual property theft is “unacceptable.”
• Chinese AI companies could face sanctions if wrongdoing is established.
• The administration is evaluating enforcement options.

Policy Overview

Item	Details
Country	United States
Official	Treasury Secretary Scott Bessent
Target	Chinese AI companies and open-source AI models
Concern	Alleged intellectual property theft
Possible Action	Economic sanctions and other restrictions

Distillation at the Center of the Dispute

The allegations focus on AI model distillation, a technique in which a smaller or newer model learns from the outputs of a more capable model. Distillation is a legitimate machine learning technique when performed with authorization, but U.S. AI companies argue that unauthorized use of proprietary frontier models can amount to intellectual property theft.

American AI firms, including Anthropic, have previously accused some Chinese developers of using their models to accelerate development of competing systems. Reports indicate that companies such as Moonshot AI, MiniMax, and DeepSeek have come under scrutiny in Washington, although none has been formally found liable for IP theft.

Kimi K3 Draws Washington’s Attention

The debate has intensified following the emergence of Moonshot AI’s Kimi K3, an open-source model that has demonstrated performance comparable to several leading Western AI systems.

According to reports, U.S. officials are examining whether advanced Chinese models incorporate techniques or knowledge derived from proprietary American AI systems. Some officials have also referenced digital “watermarks” and other forensic methods that could help identify unauthorized model reuse, although no formal findings have been announced.

Why Kimi K3 Matters

Factor	Significance
Open-source release	Broad global accessibility
High performance	Narrows capability gap with U.S. models
Policy impact	Intensifies debate over AI IP protection
Regulatory response	Could influence future U.S. sanctions

Broader US-China AI Competition

The sanctions warning comes amid an increasingly tense technological relationship between the world’s two largest economies.

Recent U.S. actions have included:

• Tightening AI chip export controls.
• Restricting access to certain advanced American AI models.
• Increasing scrutiny of Chinese AI development.
• Exploring new safeguards for frontier AI technologies.

Meanwhile, China is reportedly considering tighter controls on exports of AI technologies and access to some advanced domestic AI models, underscoring the growing strategic importance of artificial intelligence.

Industry Impact

If sanctions are ultimately imposed, they could have significant consequences for the global AI ecosystem.

Potential effects include:

• Reduced international adoption of targeted Chinese AI models.
• Increased compliance requirements for companies using foreign AI systems.
• Further fragmentation of global AI development.
• Stronger emphasis on protecting AI intellectual property.
• Accelerating the split between U.S. and Chinese AI ecosystems.

Looking Ahead

The Trump administration’s warning represents another step in the intensifying competition between the United States and China over artificial intelligence leadership. Although no sanctions have yet been announced, Washington’s willingness to consider punitive measures signals that intellectual property protection is becoming a central element of U.S. AI policy. Whether the government proceeds with sanctions will likely depend on its ability to substantiate claims that Chinese AI models were built using proprietary American technologies without authorization.

The dispute also highlights a broader challenge facing the AI industry: balancing the rapid spread of open-source innovation with the protection of intellectual property. As frontier AI models become more powerful and globally accessible, governments are expected to devote greater attention to model provenance, responsible AI development, and cross-border technology governance, making AI a key battleground in the evolving U.S.-China technology rivalry.