AI as a question of national security. Chinese authorities held consultations with the country’s leading technology companies in June about potentially restricting foreign access to China’s most advanced AI models. The Reuters news agency reported this on Tuesday with reference to three people familiar with the discussions.
Read more after the ad
The talks, which included representatives from Alibaba, ByteDance and the AI start-up Zhipu AI (known as Z.ai), underscored, according to Reuters, that Beijing, like Washington, now views advanced artificial intelligence (AI) as a critical national asset that requires state control. The meetings, led by China’s Ministry of Commerce, discussed restrictions on China’s most advanced AI models – both closed-source and more open versions, according to two of the Reuters sources.
Theft of AI technology as a criminal offense
All three companies mentioned have different AI models. Some of these are closed-source, meaning closed models, while others are open-weight models, meaning users can download, run and customize the underlying systems. The most widely used AI models in China include Alibaba’s Qwen and ByteDance’s Doubao. Z.ai has recently made waves in Silicon Valley because the performance of its GLM-5.2 model approaches leading US models like Anthropic’s Mythos – and at a fraction of the cost.
According to one of the sources, the meetings also discussed making leaking or stealing proprietary AI technology a criminal offense under the national security law. The government representatives also spoke about possible new regulations regarding who can finance domestic AI start-ups.
The scope of the possible restrictions is still being discussed, according to Reuters. According to two of the sources, they may only be applicable to future AI models. It is also unclear when or whether these measures will come into force at all. Neither the Chinese government nor Alibaba, ByteDance or Z.ai responded to Reuters’ queries on the matter.
Danger to national security?
The increasingly powerful AI models have raised concerns about the impact on national security, and not only in Beijing, due to their possible misuse. When the US AI company Anthropic presented its new powerful AI model Mythos, which can identify security gaps in software systems, in April there was great excitement. Myth was too powerful and therefore too dangerous for the public, Anthropic decided. In mid-June, the US government demanded a block on Anthropic’s AI models Fable and Mythos. Foreign nationals are not allowed to have access to Anthropic’s most advanced AI models, it said. Fear of espionage by China is said to have been the trigger for the bans. The US government later partially lifted the export restrictions.
Read more after the ad
For their part, Chinese authorities are deeply concerned that Mythos could exploit software vulnerabilities and that the U.S. government could use the model against Chinese interests, according to two of the Reuters sources. Recently, Alibaba banned Anthropic’s AI model Claude. For fear of surveillance, Alibaba employees are no longer allowed to use Claude.
China has also taken numerous measures this year to keep domestically developed AI technology in the country. At the end of April, the government vetoed the $2 billion takeover of the Chinese AI start-up Manus by the Facebook group Meta. As part of the review of the deal, Chinese authorities banned Manus CEO Xiao Hong and chief researcher Ji Yichao from leaving the country. In late May, the government expanded existing exit restrictions for AI talent, including AI executives from private companies such as Alibaba and DeepSeek. This year, Beijing also launched investigations into Manus and other Chinese AI startups that have moved abroad to determine whether they have violated export control laws, multiple Reuters sources reported.
New regulations and possible tiered model
A few weeks ago, Chinese authorities issued sweeping new regulations that tighten controls on foreign transactions involving Chinese investors, technology, data and national security. These could now be further tightened – this is what the consultations between the government and companies suggest.
According to its own information, Reuters was unable to find out what possible new restrictions on access to Chinese AI models from abroad could actually look like. In this context, the news agency refers to a discussion group held in May by Chinese legal experts on the regulations for open source AI. According to a summary of the discussions published in an official journal of the Supreme People’s Court, participants proposed a multi-tiered system: simple open-source AI models would only be subject to a straightforward reporting requirement, more advanced models would be subject to security reviews, and the top models would either not be made publicly available or only be used domestically.
Since the launch of DeepSeek’s R1 model last year, Chinese AI models have gained traction worldwide thanks to low costs and rapidly growing capabilities. If the Chinese government decides to significantly restrict access to Chinese AI models, this could have an impact on AI markets, as this would likely result in increasing costs for many companies, according to Reuters.
(noun)
