China weighs curbs on overseas AI model access

Chinese authorities, led by the China — Ministry of Public Security (China), have held meetings with top tech firms including Alibaba Group, ByteDance, and Z.ai over the past month to discuss potentially restricting overseas access to China's most advanced AI models, including those yet to be released. The talks follow steps by Beijing to keep homegrown AI within the country and mirror similar US moves. Officials discussed making theft of proprietary AI technology an offense under national security law and implementing new measures to restrict funding of domestic AI startups. The scope is still under discussion and may only apply to future models. The US has also limited access to its top AI models, with Anthropic restricting access to its Fable and Mythos models. China has launched investigations into Manus and other AI startups for potential export control violations. A May roundtable proposed a tiered system for open-source AI regulation.