U.S. Threatens Sanctions on Chinese AI Models Over IP Theft

U.S. Eyes Sanctions Against Chinese AI Firms

The United States is preparing to scrutinize Chinese open source AI models for evidence of intellectual property theft, with Treasury Secretary Scott Bessent warning Tuesday that sanctions could follow if violations are confirmed.

"This administration supports open source models, but what we do not support is IP theft," Bessent said during a Fox Business interview. "If we see, especially, that overseas models are stealing from our great companies, we have the ability to sanction them because of this theft."

The statement, first reported by Bloomberg, arrives as Chinese AI companies like Moonshot AI gain ground with increasingly capable models that could undermine the business prospects of leading American firms including OpenAI and Anthropic. These competitive pressures raise concerns about U.S. companies' ability to secure continued funding for frontier model development.

Why It Matters

Sanctions targeting AI models themselves would represent a major escalation beyond existing chip export controls and could reshape how open source AI development proceeds globally. The move also highlights growing tension between protecting intellectual property and fostering open collaboration in AI research—a debate that divides even U.S. technology leaders.

Distillation Debate Divides Industry

At the center of the IP theft allegations is model distillation, a technique that transfers capabilities from larger AI systems into smaller, more efficient versions. American AI labs have spent months warning about foreign actors copying their technology and releasing it as open source. The White House pledged in April to work with AI companies to combat such theft.

Yet not everyone agrees that distillation constitutes theft. Microsoft CEO Satya Nadella recently criticized major labs for this position, calling it "ironic" that companies claiming fair use rights to train on public data then "impose restrictive terms on distillation."

Hugging Face CEO Clem Delangue offered another perspective on a recent TechCrunch Equity podcast episode, arguing that distillation plays only a minor role in China's AI progress. "We know distillation to be a very small factor in the ability to create good models, and it's a practice that everyone is doing, including companies in the U.S.," Delangue said. He attributed Chinese advances instead to "really, really good research teams" taking a more collaborative approach than their American counterparts.

Legal Risks Cut Both Ways

The IP theft accusations come as U.S. AI companies face their own legal challenges over training practices. Anthropic received judicial approval this week to begin paying authors as part of a $1.5 billion settlement after a court ruled the company illegally downloaded and stored millions of copyrighted books for AI training.

Reports emerged Monday that the Trump administration is considering a complete ban on Chinese open source models, though some sources have disputed that claim. If implemented, sanctions would add another tool to Washington's strategy for maintaining AI leadership, following restrictions on advanced chip access and tightened export controls.

Details were first reported by Bloomberg and TechCrunch.