中国拟收紧开源权重模型及23项关键AI技术的出口管制

China's commerce and science ministries have released a draft update to the nation's list of banned and restricted export technologies.

The catalog targets 23 key AI technologies across large models, agents, compute clusters, and data processing. AI tools that can write or alter malware on their own to attack critical infrastructure, the military, or government agencies face a total export ban. Certain AI tools for finding software vulnerabilities won't be banned outright, though they will face new export limits.

Large AI models also face export controls in two cases. The first covers models trained on China's "important data" (as defined by the country's Data Security Law and related rules) that offer services based on it. Lawfully public data is exempt. Tiny amounts of up to 0.1 percent are also allowed if the data was cleaned beforehand, never shows up in AI outputs, and doesn't substantially boost the model's R&D capacity for core tech in that field.

The second case covers the core technologies behind large models, including weights, checkpoints, source code, architecture, and training or inference methods. These controls kick in once a model hits 10^26 FLOPs of training compute, has at least 2 trillion parameters, or ranks in the global top 10. Export officials determine that top-10 status using public rankings from major domestic and foreign evaluators or industry platforms, looking at overall performance, specific tasks, safety, or global usage numbers.

"This move does not mean 'China is shutting down open-weight models,' but it does indicate the Chinese government establishing the optionality to shut down future open-weight models at will," writes Dean W. Ball, head of strategic futures at OpenAI. Ball adds that Beijing's framework is a "mirror image" of the model-weight export controls the US government has already put in place.

This will likely push US companies that rely on open-weight AI to double down on domestic models instead of Chinese alternatives. Europe is in a much tougher spot. The EU has been banking on running open-weight models in domestic clouds to stay independent, alongside companies that want to host closed top-tier models on local servers. That sovereignty strategy falls apart when all the best open models come from the US or China and can get cut off by export controls at any time. At least AI tech that is already public through open-source releases or research papers skips these new rules entirely.

Fast-track rules protect China's open-source boom while restricting local foreign deployments

To avoid choking off its domestic open-source boom, China is introducing a fast-track process at the same time. Publishing restricted AI tech and weights as open source qualifies for "automatic licensing." The same goes for Chinese firms that run their own data centers abroad, tape out chips overseas, conduct research with foreign partners, or help set international standards. Under the ministry's rules, eligible exports are cleared the moment a company files the paperwork.

Using restricted models through API calls, apps, websites, SaaS tools, or lightweight versions on local devices stays free of these export controls. Installing restricted AI tech on local servers abroad, however, requires a standard export license. Companies can still use the streamlined process if the software runs on their own overseas cloud.

Beijing also pledges to back AI development in the Global South and ease exports for capacity building, tech sharing, and joint rulemaking within the World Artificial Intelligence Cooperation Organization, a bloc China launched to counter Western AI bodies. Until the final rules take effect, companies must register affected exports through the commerce ministry's portal.

AI News Without the Hype – Curated by Humans

Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section.