- Introduction to the diplomatic contradiction
- Promoting open sources in United Nations corridors
- Strategy to reach out to the Global South
- Secret internal discussions to impose export restrictions
- Enforcing national security laws and future implications
- Frequently asked questions
Introduction to the diplomatic contradiction
China’s technological strategy is currently experiencing a unique and complex state of clear contradiction that has caught the attention of analysts around the world. It combines international diplomatic calls to adopt open-source artificial intelligence with intensive internal moves to impose strict export controls on advanced technology. While Chinese officials were strongly promoting open-source models as a vital and effective tool for achieving comprehensive global development during official United Nations events in Geneva, Switzerland this week, reliable media reports indicated that Beijing is quietly and seriously exploring the possibility of imposing tight restrictions on external access and global markets to its most advanced language models. This stark contradiction between open diplomatic rhetoric and internal protective measures has sparked a wave of scrutiny, questions, and criticism from developers and technology policymakers alike around the world.
Promoting open sources in United Nations corridors
Chinese Minister of Industry and Information Technology Li Lecheng utilized the global UN dialogue platform on AI governance held on July 8 to clearly and loudly call for establishing “fair and inclusive” governance of artificial intelligence technologies. In his speech, he emphasized the utmost importance of enhancing the effective participation of developing countries in shaping international rules and regulations governing this vital sector. According to the official newspaper China Daily, the Chinese minister proposed launching a massive initiative comprising 200 intensive training programs focused primarily on the digital economy and artificial intelligence skills, specifically directed to support countries of the Global South over the next five years. This enthusiastic speech came during a busy week of overlapping AI-related events in Geneva, which included the AI for Good Global Summit hosted by the International Telecommunication Union from July 7 to 10 at the Palexpo conference center.
Strategy to reach out to the Global South
China’s broader stance and general diplomatic approach seek to portray open-source models offered and developed by major Chinese companies—such as DeepSeek and Qwen models backed and developed by e-commerce giant Alibaba—as vital economic tools that effectively lower the barriers to adopting artificial intelligence and its advanced applications for lower-income countries. To foster this positive image, Beijing issued an official white paper last June reaffirming its serious and ongoing efforts to establish the World Organization for Artificial Intelligence Cooperation, headquartered in Shanghai. Membership will be open to any sovereign state wishing to join, with an agenda focused clearly on bridging the global digital divide and unifying technological capabilities.
Secret internal discussions to impose export restrictions
Despite this open international rhetoric promoting cooperation, an exclusive and detailed report published by Reuters on July 7 indicated that relevant Chinese authorities, led by the Ministry of Commerce, spent the entire past month holding intense and secret meetings with major domestic technology companies such as Alibaba, ByteDance, and promising startup Zhipu AI. The main goal of these closed-door talks was to discuss ways and the possibility of limiting overseas user and developer access to the most advanced AI models owned by these leading companies. These complex discussions covered both closed-source systems and open-weight models—systems that can be downloaded for free from the internet and have been the primary and core reason for the massive popularity of Chinese AI in foreign markets and among independent developers.
Enforcing national security laws and future implications
In addition to studying access restrictions, Chinese officials discussed strict proposals to make leaks or theft of AI technology owned by domestic companies a serious criminal offense punishable by law under China’s National Security Law. They also proposed specific and strict financial limitations on entities and individuals permitted to fund domestic startups operating in this sensitive field. The idea of a tiered system was raised in an official magazine affiliated with the Supreme People’s Court, subjecting basic open-source tools to simple preservation and registration procedures, while more advanced technologies face precise and complex security reviews, and leading and most sensitive models are completely banned from public release or restricted to domestic use only. Tech publication The Next Web noted this curious paradox, pointing out that China has earned global goodwill by giving away its technological models for free, as European developers have heavily relied on cheap open weights as an alternative to costly American models. This raises a fundamental question about Beijing’s ability to maintain its image as a generous benefactor to the Global South if it proceeds to restrict those models that earned it this prestigious international reputation.
Frequently asked questions
Question: What are the most prominent features of China’s diplomatic stance at recent UN conferences?
Answer: China called for establishing fair and inclusive AI governance and proposed launching 200 specialized training programs targeted at the Global South to enhance comprehensive digital development over the next five years.
Question: What restrictions and procedures are Chinese authorities currently studying in secret?
Answer: China’s Ministry of Commerce is studying restrictions on foreign entities and users accessing advanced AI models, whether closed or open-source, alongside criminalizing technology leaks and enforcing the national security law.
Question: How do developers outside China, especially in Europe, rely on Chinese technologies?
Answer: Many foreign and European developers heavily rely on cheap, free open-source models provided by Chinese companies as a practical and economical alternative to closed, highly costly American models.
Question: Will the proposed Chinese restrictions apply to models currently available to the public?
Answer: Informed media sources indicate that any new regulatory or restrictive measures will likely apply only to future models and will not retroactively affect models currently available to the public, although there is no specific launch schedule.
