- Beijing is looking at curbing overseas access to China's top AI models, sources say.
- The meetings, led by China's Ministry of Commerce, discussed limits on both closed-source and open versions of AI models.
- Officials proposed making leak or theft of AI technology an offence under national security law.
- Possible new measures to restrict who can fund domestic AI startups were raised.
- Chinese authorities are deeply worried about Mythos potentially being used by Washington against Chinese interests.
- The companies have both closed-source and open-weight models, including widely used Alibaba Qwen and ByteDance Doubao, while Z.ai's GLM-5.2 has drawn attention in Silicon Valley.
- The U.S. has also restricted access to advanced AI models, ordering that foreign nationals not have access to Anthropic's Fable and Mythos models, and Chinese authorities are worried about Mythos potentially being used against Chinese interests.
Beijing is actively exploring measures to restrict overseas access to its top artificial intelligence models, reflecting a growing concern over national security and technology control.
Chinese authorities have held meetings with major tech firms, including Alibaba and ByteDance, to discuss potential restrictions on advanced AI models, both closed-source and open versions. According to sources, the discussions are led by China's Ministry of Commerce and include proposals to make any leak or theft of proprietary AI technology an offense under stringent national security laws.12
The scope of these potential restrictions is still under discussion, with indications that they may apply only to future models. Officials are particularly worried about the capabilities of models like Mythos, fearing they could be exploited against Chinese interests.5910
In recent months, Beijing has implemented several measures to protect its homegrown AI, including ordering Meta to unwind its $2 billion acquisition of the Chinese-founded AI startup Manus. In June, new rules were introduced to tighten control over overseas deals involving Chinese technology and data.
The proposed measures may include a tiered system for AI technologies, where basic open-source tools would require simple filing, while more advanced technologies would face security reviews, and the most sensitive models would be restricted to domestic use only.
As the global AI landscape evolves, China's approach underscores its commitment to safeguarding its technological advancements and maintaining control over its AI assets.
“The talks, led by China's Ministry of Commerce, included discussions on limiting both closed-source and open-weight models, as well as making the leak of proprietary AI technology an offence under national security law. Chinese authorities also raised concerns about U.S. AI model Mythos potentially being used against Chinese interests.”
Yahoo Finance
