Post by : Badri Ariffin

Chinese authorities have held discussions over the past month with some of the country's biggest technology companies about limiting overseas access to China's most advanced artificial intelligence (AI) models, including models that have not yet been released.

According to three people familiar with the meetings, officials met with representatives from Alibaba, ByteDance, and AI startup Z.ai. The sources spoke on condition of anonymity because they were not authorised to speak publicly.

The discussions come as China continues to strengthen efforts to keep its most advanced AI technologies within the country. The move highlights Beijing's growing view that cutting-edge AI is a strategic national asset that requires tighter protection, similar to the approach taken by the United States.

Possible Restrictions on Future AI Models

During the meetings, led by China's Ministry of Commerce, officials discussed introducing limits on the country's most advanced AI models. The proposals covered both closed-source models and more open versions, according to two of the sources.

The discussions are still ongoing, and the exact scope of the restrictions has not been finalised. Two sources said any new rules may apply only to future AI models. It also remains unclear when, or if, the measures will officially take effect.

China's Ministry of Commerce and the National Development and Reform Commission, whose officials also attended the meetings, did not respond to Reuters' requests for comment.

Alibaba, ByteDance, and Z.ai also declined to comment.

Tougher Penalties for AI Technology Theft Under Discussion

Officials also discussed making the theft or unauthorised leakage of proprietary AI technology an offence under China's national security law, according to one source.

The same source said authorities also raised the possibility of introducing new measures that would limit who can financially support domestic AI startups.

These proposals remain under discussion, and no final decisions have been announced.

China's AI Models Gain Global Attention

Chinese AI models have expanded rapidly in international markets since the release of DeepSeek's R1 model last year. Their lower costs and improving performance have helped them gain wider adoption.

Industry observers believe that if China limits overseas access to these advanced models, businesses outside the country could face higher AI-related costs due to reduced availability of low-cost alternatives.

Alibaba's Qwen and ByteDance's Doubao are currently among the most widely used AI models in China.

Meanwhile, startup Z.ai has attracted attention in Silicon Valley after its GLM-5.2 model demonstrated capabilities close to leading U.S. AI systems while operating at a much lower cost.

AI Security Concerns Mirror US Approach

The discussions in China come as the United States has also increased security measures around advanced AI technologies.

The administration of U.S. President Donald Trump has expressed concerns that advanced American AI systems could be used by military or intelligence organisations in countries such as China and Russia.

In June, the U.S. restricted foreign nationals from accessing Anthropic's most advanced AI models, Fable and Mythos. Because user nationality could not be verified instantly, Anthropic temporarily disabled access to the models for all users worldwide.

After additional safeguards were introduced, export controls on Fable were removed. However, Mythos remains available only to selected trusted organisations in the United States because it is designed for cybersecurity professionals.

Some U.S. AI experts have also argued that the country should regulate the use of Chinese AI models.

China Raises Concerns Over Anthropic's Mythos Model

According to two of the sources, Chinese authorities are particularly concerned that Anthropic's Mythos model could identify software vulnerabilities and potentially be used by Washington against Chinese interests.

Similar concerns have been expressed publicly by Chinese state media and Zhou Hongyi, founder of cybersecurity company 360. Zhou has said China should develop its own AI model with capabilities similar to Mythos.

China Continues Tightening AI Controls

China has introduced several measures this year to strengthen oversight of its AI industry.

In April, the country's state planner ordered Meta to unwind its $2 billion acquisition of Chinese-founded AI startup Manus.

In early June, Chinese authorities also introduced broader regulations tightening oversight of overseas deals involving Chinese investors, technology, data, and national security.

According to two sources and a third person familiar with the matter, authorities have also investigated Manus and several other Chinese AI startups that relocated overseas to determine whether they violated export control laws.

Manus did not respond to Reuters' request for comment.

Proposed Tiered System for AI Models

Reuters was unable to determine how any future restrictions on overseas access to Chinese AI models would be implemented.

However, possible approaches were discussed during a May roundtable involving Chinese legal experts on open-source AI regulations.

According to a summary published in an official journal of China's Supreme People's Court, experts proposed a tiered system that would classify AI models based on their sensitivity.

Under the proposal, basic open-source AI tools would require only a simple filing process. More advanced technologies would undergo security reviews, while the most sensitive frontier AI models could be prohibited from public release or limited for domestic use only.

#Tech News #Global Updates #AI Technology #World News