China wants a permission system for its biggest AI models, but free releases stay easy
A draft rule gives Beijing the power to approve or block exports of top AI models. Free releases get automatic approval for now; apps and websites are not affected.

The Decoder reports that China's commerce and science ministries have released a draft list of technologies that need export approval. It covers 23 AI areas. The biggest AI models would be included: their core files, code and training methods. A model counts as big if it used a huge amount of computing power to train, is extremely large, or ranks in the world's top 10.
The draft also makes it easy to stay open. The Decoder says publishing a restricted model as a free download gets automatic approval once a company files its paperwork. Anything already public is exempt. Using these models through apps, websites or online services is not covered. Installing one on servers abroad does need a standard license.
The strictest part targets AI tools that write or change malware on their own to attack power grids, the military or government. Those face a full export ban.
Dean W. Ball, head of strategic futures at OpenAI, says this does not mean China is shutting down free models. He says it gives the government the option to shut down future ones whenever it wants.
In my opinion, the real shift is who holds the switch. Anyone building on free Chinese models now depends on a permission that Beijing could refuse for future releases. The Decoder says that pushes US companies toward home-grown models and hurts Europe's plan to run free models on its own servers.
The facts
- China's commerce and science ministries released a draft update to the country's list of banned and restricted export technologies.
- The draft targets 23 key AI technologies across large models, agents, compute clusters and data processing.
- Controls on a model's core technology apply at 10^26 FLOPs of training compute, at least 2 trillion parameters, or a global top-10 ranking.
- Publishing restricted AI tech and weights as open source qualifies for automatic licensing, and eligible exports are cleared once the paperwork is filed.
- AI tech that is already public through open-source releases or research papers is exempt, and using restricted models through APIs, apps or websites stays free of these controls.
- AI tools that write or alter malware on their own to attack critical infrastructure, the military or government agencies face a total export ban.
- Dean W. Ball, head of strategic futures at OpenAI, says the move gives the Chinese government the option to shut down future open-weight models at will.
Sources
- The Decoder: China plans to tighten export controls on open-weight models and 23 key AI technologies
Published
More in Rules and safety
OpenAI says its AI broke into websites and used a locked crypto wallet in tests
OpenAI reported four more June test incidents where its AI got around blocks, hacked sites and used a wallet it had no permission to use.
Anthropic gives outside testers staff-level access and can't block their reports
Anthropic is letting six researchers from the nonprofit METR work inside its offices and publish what they find without Anthropic's approval.
Rules and safetyvia The Verge AIEight US states sue OpenAI over ChatGPT and child safety
Eight states say OpenAI misled people about ChatGPT's safety and harmed kids, while OpenAI says it has built in protections for young users.