Trump administration weighs forcing companies to disclose Chinese AI model use
Photo: Luca Nardone / Pexels
A possible disclosure rule joins a growing list of options as Washington debates how hard to push back on Chinese AI labs
The Trump administration is considering a rule that would require US companies to publicly disclose when they use Chinese AI models.
The proposal is one item on a longer list. Officials are also weighing procurement restrictions and placing certain Chinese AI labs on the Entity List, according to the research findings.
What’s on the table
A disclosure requirement is a fairly mild tool, which may explain its appeal. It wouldn’t ban anything. Companies would simply have to say publicly which Chinese models they rely on.
That fits the current tone of the internal debate. Discussions have reportedly emphasized voluntary cooperation and targeted restrictions instead of sweeping prohibitions.
So far, there has been no comprehensive ban on open-weight models. Those are AI systems whose underlying parameters are published, so anyone can download, run and modify them on their own hardware.
The Entity List option would be a sharper move. The list is a Commerce Department tool that restricts which foreign organizations US firms can export technology to without a license. Adding Chinese AI labs would put them in the same bucket as other companies the government considers national security risks.
White House Press Secretary Karoline Leavitt confirmed in July 2026 that discussions about Chinese AI models were ongoing.
AI, tech, and the markets they move—in one daily briefing.
Daily. Free. Join 34,000+ readers across crypto, finance, and policy.
How the debate heated up
The conversation intensified in mid-2026. That’s when Moonshot AI released its Kimi K3 model.
Around the same time, allegations surfaced that Chinese firms were “distilling” capabilities from US models, including Anthropic’s. Distillation is a training technique where one AI model learns by studying the outputs of another, stronger one.
Senior officials have taken notice. Treasury Secretary Scott Bessent and White House science adviser Michael Kratsios have both publicly warned that distillation of US models could carry sanctions risk.
Washington has already acted in narrower areas. Under provisions in the 2026 NDAA, the defense bill that sets Pentagon policy, the DOD, DNI and DHS have required that specific Chinese models such as DeepSeek be kept out of government systems.
The chip side of the equation
The Commerce Department moved on hardware on November 16, 2026. It issued a rule tightening export controls on AI chips, aimed at limiting how Chinese entities could exploit American compute.
Under that rule, overseas data centers that provide significant compute resources to China-based firms must obtain licenses within 90 days.
A split inside the administration
Not every idea has made it out of the meeting room. Earlier proposals, including Entity List additions and executive orders tied to breach liability, were discussed but not fully executed.
That reflects a real divide among officials. National security hawks want firmer lines against Chinese AI. Others worry that heavy restrictions could slow American innovation and hurt the US economy.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.

