China drafts export controls on 23 AI technologies, targeting open-weight models and compute thresholds

China's commerce and science ministries have released a draft update restricting 23 key AI technologies. The rules impose export bans on AI tools capable of autonomous malware creation targeting critical infrastructure, and export limits on vulnerability discovery tools. Large AI models face controls if trained on restricted data, if they exceed 10^26 FLOPs training compute, have 2 trillion parameters, or rank in the global top 10. A fast-track process allows automatic licensing for open-source releases and overseas data center operations, aiming to protect China's domestic open-source ecosystem while restricting foreign deployments.