Why Chinese military researchers are using US AI models
What's the story
Chinese military researchers have been using outputs from leading US artificial intelligence (AI) models developed by OpenAI and Anthropic. The information was revealed after a Reuters review of over 80 Chinese academic papers and patents. This practice is part of China's efforts to improve its defense capabilities, despite Washington's attempts to restrict Beijing's access to advanced chips and other strategic technologies.
Technique
Widespread use of 'model distillation'
The documents reviewed by Reuters show a widespread use of "model distillation."
This technique involves using outputs from powerful AI systems to train smaller, specialized models that can be deployed locally without the huge computing power needed to build frontier AI systems from scratch.
The review revealed that this method is widely used by researchers associated with the People's Liberation Army and other military institutions in China.
Strategic perspective
Major flashpoint ahead of US-China talks
The research papers suggest that Chinese defense institutions view leading US AI models as a source of technical insight and a way to close the gap with American rivals.
This has become a major flashpoint ahead of US-China talks on AI governance and safety.
US officials have accused some Chinese entities of using distillation to extract capabilities from American AI models, potentially undermining export controls and infringing intellectual property rights.
Rebuttal
China rejects allegations, accuses US of AI 'hegemonism'
China has rejected the allegations, accusing Washington of AI "hegemonism" while arguing that US firms have engaged in similar practices.
Chinese developers have also disputed claims that their AI advances rely on foreign models.
For instance, AI start-up Moonshot recently denied allegations by the Donald Trump administration that its Kimi K3 model was built using distillation, asserting it was driven by proprietary innovations.
Adaptation
Transferring proprietary reasoning from Western models into smaller systems
Sunny Cheung, a Jamestown fellow who analyzed over 60 of the papers, said Chinese military scientists are systematically capturing the reasoning steps of Western models to adapt them for surveillance, cyber warfare, and tactical decision-making.
"Teaching a model the right answer is one thing but teaching it the reasoning behind the answer is much harder," said Cheung.
He added these papers show researchers are trying to transfer expensive, proprietary reasoning from Western models into smaller systems they can deploy locally.
Diverse applications
Content monitoring to military deployment
Chinese researchers have employed distillation for a range of purposes, from content monitoring to military deployment.
At the North University of China, closely linked to the country's weapons industry, researchers used Anthropic's Claude 3 Haiku to generate synthetic training data for a text classification model for social media monitoring and content moderation.
However, Anthropic said it does not provide commercial access to Claude in China or Beijing-controlled firms and uses monitoring systems to detect policy violations.
Competitive edge
Competing with the US in frontier AI
China has adopted distillation as a way to compete with the US in frontier AI, despite facing restrictions on advanced computing resources due to Washington's export controls on high-end chips.
Central and local governments have promoted "model lightweighting" and edge computing, directing subsidies and research funding toward technologies that enable AI models to run on drones, satellites, and other devices with limited processing power.