AI export controls may soon reach beyond the hardware used to run AI. China is proposing rules that would also cover advanced models and AI systems that can act with little human direction.
The draft would add AI-related technologies to China's list of exports that are banned or restricted. It also covers technology for designing chips and operating large computing facilities.
The important distinction is between using a Chinese AI service and taking its underlying technology abroad. Overseas users could generally still access services through websites and cloud platforms. But transferring models, data or expertise, and some overseas deployments, would need government approval. In this proposal, AI policy reaches beyond who can use a tool to who can move the technology behind it across borders.
Think of it as letting someone use a service while controlling whether they can take its recipe and expertise across a border.
For companies working across countries, access to an AI service may remain available while moving the technology behind it becomes a regulated step.
For now, these are proposed rules, not binding restrictions. The draft remains open for public comment until November 15, 2026. The question is whether AI systems and know-how become a regular part of export-control policy worldwide, and how companies adapt if they do.
