A recent report reveals that researchers connected to the Chinese military, including the People's Liberation Army (PLA) and security institutions, are allegedly leveraging outputs from advanced US artificial intelligence models, specifically those developed by OpenAI and Anthropic. This practice is reportedly being used to train China's own AI systems for various defence and military applications.
Unauthorized Use of US AI Models
According to a Reuters investigation, which reviewed over 80 Chinese academic papers and patents, Chinese researchers are employing a technique known as “model distillation.” This method involves using powerful, existing AI models to teach a smaller, domestic model to replicate specific capabilities. This approach allows organizations to develop sophisticated AI systems more rapidly and with fewer computational resources, bypassing the need to build everything from scratch.
The US government has been actively working to limit China's access to advanced technologies, including high-end AI chips, due to national security concerns. OpenAI and Anthropic, the companies whose models are allegedly being used, do not authorize access to their services within China. This unauthorized use raises significant concerns about intellectual property theft and the uncontrolled global dissemination of sensitive AI knowledge.
Applications and Implications for Defence
The distilled AI systems developed by these Chinese military-linked researchers are reportedly being explored for a wide array of defence applications. These include enhancing drone operations, executing cyber warfare-related tasks, monitoring social media, and improving capabilities in maritime and battlefield operations.
Sunny Cheung, a researcher at the Jamestown Foundation, highlighted the complexity of this transfer. "Teaching a model the right answer is one thing, but teaching it the reasoning behind the answer is much harder," Cheung stated. "These papers show Chinese military-linked researchers are trying to transfer that expensive, proprietary reasoning from Western models into smaller systems they can control and deploy locally."
The alleged transfer of internal reasoning abilities from Western AI models into Chinese military systems represents a significant development. It suggests a strategic effort to acquire advanced AI capabilities without the extensive investment in foundational research and development, potentially accelerating China's military modernization efforts.
International Response and Future Restrictions
The US government is currently discussing potential restrictions in response to these allegations. While China has rejected US claims, accusing Washington of pursuing AI “hegemonism,” the controversy underscores the growing geopolitical tensions surrounding AI development and its military implications. The unauthorized access and use of these advanced AI models could trigger broader national security risks and complicate international efforts to control the spread of critical AI technologies.