The Trump administration is weighing a requirement that would force US companies to disclose whether they use Chinese artificial intelligence models, according to a Reuters exclusive published Tuesday, the latest sign that Washington is broadening its response to China's advances in AI beyond chips.
The Treasury Department and the White House Office of Science and Technology Policy have discussed the plan, Reuters reported, citing multiple people familiar with the matter. Under the proposal, US firms that build products or services on Chinese-developed AI models would have to say so publicly — an approach that has been described in press coverage as a "name and shame" strategy designed to discourage adoption without outright bans. For more context on this story, see our ongoing artificial intelligence updates.
Why Disclosure, Why Now
The deliberations reflect how quickly Chinese AI models have spread through the US technology ecosystem. Chinese developers have widely adopted an open-weight approach, releasing the core components of their models so that anyone can download, modify and deploy them. US companies have increasingly taken advantage, citing low costs and strong performance, Reuters reported.
The most prominent example is Airbnb. According to Reuters, the home-sharing company uses Alibaba's Qwen model — and if the disclosure requirement takes effect, companies like Airbnb would have to state on their websites that their products rely on Chinese AI models.
Alibaba's Qwen family has become one of the most downloaded open-weight model lines in the world, and DeepSeek, Moonshot AI and other Chinese labs have released competitive open models that US startups, researchers and enterprises routinely deploy. That uptake has quietly built a layer of Chinese technology into American products — the kind of reliance that has historically triggered scrutiny in Washington when it involved telecoms equipment or social media apps.
Disclosure Instead of Bans
The administration has so far stopped short of restricting Chinese AI models outright or sanctioning the companies behind them. A disclosure mandate would apply pressure indirectly: by making Chinese AI use visible to consumers, business partners and investors, the policy aims to make adoption costlier in reputation and procurement terms without triggering the kind of market disruption a ban would cause.
It is also a notable shift in emphasis. US policy toward Chinese AI has so far concentrated on the supply side — export controls that keep advanced Nvidia chips out of Chinese hands, and investment rules that limit Chinese access to US capital. A disclosure rule would target the demand side, addressing the growing footprint of Chinese AI inside US companies rather than the flow of technology into China.
Lawmakers' Concerns
The push reflects mounting concern in Congress about dependence on Chinese models. Lawmakers have warned that deeper reliance on Chinese AI could undermine US technological competitiveness and create national security risks — arguments that echo the debate that preceded restrictions on Huawei's telecoms equipment and the forced divestment of TikTok.
Reuters reported that it remains unclear how much support the idea has within the administration, and no formal proposal or timeline has been announced. Any eventual rule would also face practical questions the reporting leaves open: which models would be covered, whether disclosure obligations would fall on downstream developers embedding Chinese models in their products, and how compliance would be verified for open-weight systems that can be modified beyond recognition.
What to Watch
If the administration moves forward, the disclosure requirement would land at a sensitive moment in US-China technology relations. The two governments have recently established channels for AI communication, and Chinese officials have sharply criticized US measures they see as containment. A mandate that effectively stigmatizes Chinese models — including open-weight releases that Beijing promotes as a soft-power tool — could complicate those channels further.
For US companies, the immediate impact is planning uncertainty. Firms building on open-weight Chinese models would face a choice between disclosing that reliance, absorbing the reputational cost, or migrating to domestic alternatives — a shift that would benefit US model providers and cloud platforms even without any formal restriction taking effect.
The deliberations are ongoing, and Reuters' sources cautioned that no decision has been made. But the direction of travel is clear: as Chinese open-weight models win adoption on price and performance, Washington is moving from restricting what China can build toward questioning what American companies are allowed to build on.
Sources
This article is based on a Reuters exclusive of November 18, 2026, corroborated by Bloomberg Law, the Seoul Economic Daily and other outlets citing the Reuters report.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →