The United States government is reportedly considering restrictions on American companies' use of foreign artificial intelligence models after a CNBC report citing an unnamed State Department source said officials see significant risks in firms deploying Chinese-developed AI.
According to the source, the principal concern is that Chinese models could amplify Beijing's narratives, suppress critical viewpoints, and promote the ideology of the Chinese Communist Party.
Why companies opt for Chinese models
The report notes that many U.S. companies choose Chinese AI models for their cost-effectiveness. Brian Armstrong, chief executive officer of Coinbase, has written that his company uses two Chinese-developed models: GLM 5.2, developed by Z.Ai, and Kimi 2.7, a product of Moonshot. A separate startup, Lindy, has said it adopted DeepSeeket to reduce company expenses.
Connection to recent U.S. actions
The CNBC piece also suggests that a desire to favor U.S.-developed AI products may be a motivating factor. The U.S. government recently halted two new models from Anthropic — Mythos 5 and Claude Fable 5 — and that action encouraged some firms to explore foreign alternatives, including Chinese models, amid uncertainty about whether similar interventions might affect other American companies in the future.
Legal and practical obstacles
It is unclear what mechanisms the U.S. government would use to force changes. Engadget has noted that restricting the use of open-source products could raise constitutional issues. Practical complications also exist: for example, Apple reportedly uses an Alibaba model in iPhones sold in China, illustrating how multinational product integration can complicate straightforward bans or limits.
Chinese discussions on outbound use restrictions
At the same time, Chinese authorities are engaged in talks with companies about limiting the overseas use of Chinese artificial intelligence, adding another dimension to the international regulatory landscape.
Conclusion
U.S. officials are weighing measures to curb American firms' use of Chinese AI models over security and political concerns, but legal, practical and economic considerations — together with Chinese regulatory intent — make implementation complex.
