There is a stir in most countries regarding maximum and correct use of AI technology. China is also working rapidly in this direction. The biggest challenge for him is to prevent advanced AI technology and models from being leaked or stolen. A report has claimed that Chinese officials held several meetings with big tech companies last month. In these meetings, there was discussion about limiting foreign access to China’s most advanced AI models.
However, the advanced AI model has not been released yet. Three people with knowledge of the conversation told news agency Reuters that Chinese officials have held a serious meeting with tech companies to prevent the leaking of China’s most advanced AI models. This conversation comes after several steps taken by Beijing to keep domestic AI within the country.
This shows that, like America, China now considers advanced Artificial Intelligence (AI) as an important national asset over which it wants to maintain control. These people said that the companies involved in the talks included tech giants Alibaba and ByteDance as well as startup Z.ai. The situation was such that these people were not even allowed to talk to the media and they refused to reveal their identities.
Since the release of DeepSeek’s R1 model last year, Chinese AI models have been spreading rapidly around the world due to their low cost and increasing capabilities. If Beijing takes any decision to limit access to these products, it is likely to have a significant impact on the AI market as costs may increase for many businesses.
Provision for strict punishment for AI theft!
According to two sources related to the matter, in these meetings led by China’s Commerce Ministry, the issue of imposing control on both the most advanced AI models (closed-source and more open versions) was discussed. A source said that officials have discussed making any leak or theft of proprietary AI technology a crime under China’s stringent national security law.
The source further said that officials also discussed the possibility of introducing new measures to restrict who can fund domestic AI startups. Two sources said the scope of possible restrictions are under discussion and may only be applied to future models. However, it was not immediately clear when these would be implemented or whether they would be implemented or not.
How will Alibaba-ByteDance react?
China’s Commerce Ministry, which oversees export regulations, and officials from the National Development and Reform Commission – the country’s government planning agency – also attended the meetings, but did not respond to questions from Reuters.
Additionally, Alibaba, ByteDance and Z.ai companies also did not respond to Reuters’ questions. All three companies have many AI models. Some of these are closed-source while some are open-source, which means that users can download, run and customize the systems inside them.
Alibaba’s Qwen and ByteDance’s Doubao are the most used AI models in China. Recently, Z.ai has created a huge stir in Silicon Valley, because the capabilities of its GLM-5.2 model are at par with America’s top models, while its cost is much lower.
Concerned about AI models and national security
The administration of US President Donald Trump is also very concerned about the impact of AI on national security. Especially regarding the fact that American AI products can be misused by the military intelligence of China, Russia and many other countries.
In June, the government ordered that foreign nationals not be given access to Anthropic’s most advanced Fable and Mythos models. After this, the company stopped the supply of these models to all the users around the world, because no one’s nationality could be confirmed in real time. Even the export control on Fable made for the general public was removed, because now new security measures have been implemented on it.
But Mythos, designed for cybersecurity professionals, is still available only to a few “trusted” US organizations. Some American AI experts have also said that America needs to regulate the use of Chinese AI models.
Concern in China about the danger of Mythos
According to two sources close to the matter, Chinese officials are very concerned that Mythos could take advantage of software vulnerabilities and that Washington could use this model against China’s interests. Government media and founder of cyber security firm 360, Zhou Hongyi, have also publicly expressed concern in this matter.
The company, 360, is a large company serving government and corporate clients, and Zhou says China needs to develop its own mythos.
Investigation into local AI startups started in China?
Meanwhile, this year China has taken many strict steps to protect the AI made in its country. In April, the country’s government planning agency ordered Meta to cancel its $2 billion acquisition of Chinese-origin AI startup Manus. Beginning in June last month, authorities implemented new and tougher rules that tightened controls on foreign deals involving Chinese investors, technology, data and national security.
According to two sources and a third person, China has started investigating Manus and other local AI startups that have moved abroad this year to determine whether they violated export control laws in any way. However, Manas has also not given any answer in this matter.
Leave a Reply