Chinese authorities have held closed-door meetings with the country’s largest technology companies to discuss possible restrictions on foreign access to their cutting-edge artificial intelligence models, even those that have yet to reach the market.
Chinese authorities have held closed-door meetings with the country’s largest technology companies to discuss possible restrictions on foreign access to their cutting-edge artificial intelligence models, even those that have yet to reach the market.
This was reported by the Reuters news agency, citing three insiders.
They report that the talks, organized by China’s Ministry of Commerce, have involved giants such as Alibaba and ByteDance, as well as promising startup Z.ai. The measures are a continuation of Beijing’s strategy to keep cutting-edge technologies at home. China is increasingly viewing AI as a key national asset that requires strict government regulation — similar to the approach taken by the United States.
One of the key issues at the meetings was strengthening accountability for technology leaks. Officials discussed the possibility of making the theft or unauthorized transfer of patented AI developments a crime under China’s national security law. They also considered limiting the number of investors who can fund Chinese startups in this field.
It is expected that future restrictions may be implemented on a multi-level principle. Legal experts suggest dividing the systems into three categories:
The rapid expansion of Chinese technology began after last year’s successful release of the R1 model from DeepSeek. Chinese developments are quickly conquering the global market due to their low cost. In particular, the GLM-5.2 model from Z.ai has caused serious concern in Silicon Valley, as it is close to American counterparts in terms of capabilities, but costs much less. Any export restrictions from Beijing could lead to increased costs for many foreign companies that rely on cheap Chinese software as an alternative to American systems.
China is also driven by defensive motives. The authorities are concerned that the United States could use new cybersecurity tools, including Anthropic’s Mythos model to detect systemic vulnerabilities, against the country’s interests. Against this background, Beijing is calling for the creation of its own defensive counterparts.
China has already taken several steps this year to protect its AI market. In April, a state planning agency blocked Meta Corp.’s $2 billion acquisition of Chinese startup Manus. In early June, the country also passed new rules that tighten oversight of deals involving foreign investors, data, and technologies that affect national security. Several local startups that have moved operations overseas are currently under investigation for violating export control laws.
It was recently revealed that ByteDance and Alibaba are rolling back features that allowed users to create and communicate with personalized AI companions. The companies are preparing to introduce strict new rules on human-AI interactions.
