China Considers Restricting Global Access to Its Advanced AI Models

China is exploring the possibility of limiting access for users outside its borders to some of its most advanced artificial intelligence models. This move could alter the international distribution of Chinese AI technologies, placing tools such as Kwin, Dubaio, and GLM under a new framework that connects technological superiority with sovereign control.

According to a Reuters report, Chinese authorities have held discussions with major companies including Alibaba, ByteDance, and Z.ai. These talks focused on potentially restricting external access to the most sophisticated AI models, including those not yet released. The conversations addressed both closed models and certain open-source versions, considering the handling of leaks or theft of sensitive AI technologies as possible violations of national security laws.

Reuters noted that Chinese AI models gained global traction following the rise of DeepSeek, benefiting from lower costs and improved performance. However, any new restrictions could increase expenses for international companies that rely on these models as more affordable alternatives to American AI solutions. The discussions also include the possibility of imposing limits on foreign funding for local startups, introducing a financial dimension to the technical restrictions. This approach would position Chinese AI closer to a strategic, closed dossier rather than a globally accessible product.