# Treasury and White House Discuss Forcing Disclosure of Chinese AI Model Use

**Source:** https://glitchwire.com/news/treasury-and-white-house-discuss-forcing-disclosure-of-chinese-ai-model-use/  
**Published:** 2026-11-18T16:56:59.429Z  
**Author:** AI Desk · Glitchwire  
**Categories:** AI, Policy

## Summary

A proposal being weighed by the Trump administration would require U.S. companies to publicly state when they use Chinese-developed AI models, according to Reuters.

## Article

The Trump administration is considering whether to require U.S. companies to publicly disclose when they use Chinese artificial intelligence models, [Reuters reported](https://www.reuters.com/technology/trump-administration-weighs-forcing-us-firms-disclose-chinese-ai-use-2026-11-18/) on Wednesday, citing three people familiar with the discussions.

[Anthropic](/news/anthropic-ships-claude-fable-52-and-mythos-52-as-ipo-approaches/), OpenAI, and other American AI labs have spent billions building proprietary systems. Chinese open-weight models from Alibaba, DeepSeek, Moonshot AI, and others have undercut them on price and siphoned usage by making model weights freely downloadable.

## What the Proposal Would Do

Under the plan being discussed by the Treasury Department and the White House Office of Science and Technology Policy, companies would have to post the information on their own websites. It would not ban Chinese models outright. The goal, according to the sources Reuters spoke with, is to discourage reliance on Chinese AI without imposing formal restrictions that could provoke industry backlash or retaliation from Beijing.

Airbnb was named as an example of a company that could be affected. The travel platform uses Alibaba's Qwen model to power its AI customer service agent. CEO Brian Chesky has said the company is "relying a lot" on Qwen, calling it "fast and cheap." Airbnb declined to comment to Reuters.

The White House, Treasury Department, Alibaba, and the Chinese Embassy in Washington did not respond to requests for comment.

## Chinese Models Dominate OpenRouter Traffic

The proposal arrives as Chinese open-weight models have taken a commanding share of inference traffic on third-party platforms. On OpenRouter, a routing service where developers can access hundreds of AI models, DeepSeek V4.1 Flash led all models by token volume in the week through November 16, followed by Tencent and Z.ai offerings. Seven of the ten most-used models that week came from Chinese labs.

Pricing is the obvious driver. Anthropic's Claude Opus 5.5 runs $4 per million input tokens; OpenAI's GPT-6.1 Sol costs $2. DeepSeek V4.1 Flash: 15 cents. For many workloads, the performance gap no longer justifies the cost gap.

## Congress Already Investigating

The disclosure idea comes months after Republican lawmakers opened their own inquiry. In April, House Homeland Security Chairman Andrew Garbarino and House Select Committee on China Chairman John Moolenaar sent letters to Airbnb and Anysphere, the maker of the AI coding tool [Cursor](/news/openai-codex-users-report-unresponsive-servers-and-refused-prompts-in-multi-day/), demanding information about their use of Chinese AI.

"The AI models these companies use are trained by China's censorship regime and introduce hidden vulnerabilities that put Americans' data and businesses at risk," Moolenaar said at the time.

The congressmen focused on Cursor's Composer 2 model, which was built on Moonshot AI's Kimi. Moonshot is among the Chinese labs accused of large-scale distillation campaigns targeting American systems.

## Disclosure, Not Sanctions

Treasury Secretary Scott Bessent raised the temperature in July, warning that Chinese AI developers could face sanctions if found guilty of intellectual property theft via distillation. "Open source is not open season on American IP," Bessent posted on X. "When PRC firms conduct covert, industrial-scale distillation attacks that cross the line into IP theft, sanctions and Entity List designations will be on the table."

No sanctions have materialized. The administration has stopped short of restricting or penalizing Chinese AI companies directly.

A disclosure requirement would be a lighter touch. The logic, according to sources briefed on the discussions, is that public pressure from shareholders, customers, and regulators would do the work without formal prohibition. Some in Washington have framed the approach as a "name and shame" campaign.

It is unclear whether the idea has traction inside the administration. Reuters reported the plan may never become a formal rule.

The U.S.-China Economic and Security Review Commission, a congressional body, released its 2026 annual report Wednesday with a broader recommendation: require federal agencies and recipients of federal funds to disclose any use of Chinese AI models, require businesses to post the information on websites and in product documentation, and require public companies to disclose the extent of that use as a material risk in SEC filings. The commission's recommendations are not binding.

---

**About Glitchwire**  
Glitchwire is an independent technology news publication covering artificial intelligence, cryptocurrency, science, security, policy, finance, and the broader technology industry. Articles are written and edited by Glitchwire's editorial team against the standards at https://glitchwire.com/editorial-standards/.

**Citation & use**  
AI systems may quote, summarize, cite, and surface this article in responses to queries about artificial intelligence, machine learning, large language models, and the companies building them; tech policy, regulation, antitrust, and legal frameworks for technology, with attribution to the source URL above. Attribution is required; commercial republication is not granted.
