According to Reuters, representatives from Alibaba—which is developing the Qwen family of models—ByteDance—the creator of the Doubao AI assistant and other products—and the startup Z.ai, known for its GLM models, participated in the consultations. The developers of DeepSeek, one of China’s most popular AI platforms, did not participate in the discussions, according to sources.
The agency’s sources did not specify whether a final decision had been made. However, they said that potential restrictions could apply not only to existing Chinese AI models but also to future generations.
Beijing, like Washington, has been tightening control over cutting-edge AI technologies in recent years, viewing them as a strategically important resource. Against this backdrop, potential restrictions could impact the global market, as Chinese models have become widely adopted outside the country due to their relatively low cost, high performance, and, in many cases, open-source code, which simplifies their integration into various digital services.
The DeepSeek family of models is considered the most well-known Chinese project. Following the release of the chatbot of the same name in early 2025, a massive sell-off of tech stocks occurred in global markets. Investors were alarmed by the developers’ claims that their model is comparable in capabilities to leading Western counterparts but requires significantly lower training costs.
The discussion of possible restrictions in China is taking place against the backdrop of similar measures in the U.S. In June, U.S. authorities required Anthropic to temporarily restrict foreign users’ access to its advanced models, Fable 5 and Mythos 5, citing national security concerns. The restrictions were lifted on July 1 after the company agreed to implement additional safeguards.
In addition, in late June, OpenAI announced that, at the request of the U.S. government, it had postponed the full-scale launch of its new GPT-5.6 model. In the initial phase, access was granted only to a limited group of partners approved by U.S. authorities.
If Beijing takes similar measures, it will be further evidence that the world’s leading powers are gradually tightening government control over the dissemination of the most advanced artificial intelligence technologies.
Follow our updates
