344
Productivity & Workflow355
Automation & Workflow225
Software Development251
Marketing & Growth193
AI Infrastructure & MLOps176
Writing & Content Creation204
Data & Analytics142
Photography & Imaging156
Design & Creative170
Customer Support133
Sales & Outreach125
Education & Learning131
Voice & Speech135
Operations & Admin87
A new draft would restrict or require licenses for certain AI technologies, while generally leaving public code and online AI services outside the controls.
In short: China has proposed new export rules that would restrict or require licenses for some advanced AI technologies, while still allowing many companies to sell AI services overseas.
China released draft updates to its export control rules. The list includes 40 technologies across different sectors, with 23 areas related to AI. The goal is not a total ban, but a system where certain exports may need approval first.
The draft points to several AI areas, including advanced AI models, AI agents (software that can take actions for you, like a digital assistant that can carry out tasks), large computing clusters (groups of computers working together), and some chip design know-how. It could apply to AI models trained on sensitive Chinese data, models with at least 2 trillion parameters (parameters are the internal settings that help a model make answers), or models that reach certain top performance levels.
Chinese officials say the controls are meant to reduce misuse, protect national security and critical infrastructure, and keep AI under human control. A temporary registration requirement for certain transfers took effect on October 16, including some transactions already in progress. Public comments are scheduled through November 15, and final rules have not been announced.
These draft rules show a two-sided approach. China is trying to keep its most sensitive AI capabilities from being transferred abroad, while still keeping paths open for Chinese firms to compete internationally.
One key detail is that the draft generally excludes technology that is already public, such as open-source releases or published research. It also reportedly distinguishes exporting technology from letting overseas users access AI through APIs, apps, websites, or software subscriptions, which can be like letting someone use a tool without handing them the blueprint. Analysts warn that even targeted controls could add short-term costs and paperwork for companies expanding overseas.
Source: NYTimes