Now Reading: Chinese military using US AI models via distillation, reports say
-
01
Chinese military using US AI models via distillation, reports say
Chinese military using US AI models via distillation, reports say
Chinese military researchers are using outputs from leading U.S. artificial intelligence models developed by OpenAI and Anthropic to train domestic defence systems, according to a Reuters review of more than 80 Chinese academic papers and patents shared with the Jamestown Foundation. The papers show widespread use of 'model distillation', where a powerful AI system's outputs are used to train smaller, specialised models deployable locally without the enormous computing requirements of frontier systems.
Separately, a U.S. congressional advisory body has warned that China is commercialising and monetising data as a strategic national asset to fuel AI and technology goals, potentially outpacing the U.S. The US-China Economic and Security Review Commission report says China is systematically collecting domestic enterprise and physical-world data crucial for training AI tools for business, autonomous vehicles, and humanoid robots. China has rejected accusations of unauthorised distillation, calling them U.S. 'AI hegemonism'. A U.S.-China dialogue on AI governance and safety is expected.
Indian Opinion Analysis
These two sources approach the China AI story from entirely different angles, producing coverage that is complementary rather than contradictory. The Hindu, drawing on a Reuters investigation with the Jamestown Foundation, focuses narrowly on technical evidence: Chinese military researchers using model distillation from U.S. AI systems like OpenAI's GPT-3.5 and Anthropic's Claude 3 Haiku for surveillance, cyber warfare and drone targeting. The framing is neutral-report, presenting verified academic papers and official denials without editorialising. Times Now, by contrast, reports a separate U.S. commission report on China's broader data strategy, framing data as a 'strategic national asset' that could give China an edge in embodied AI for commercial and military robotics. Its stance is neutral-report but adopts the commission's warning language. Together they show two distinct prongs of U.S. concern: unauthorised extraction of Western AI models and systematic domestic data consolidation. Neither source provides a Chinese government rebuttal in detail. A careful reader should note that both reports are U.S.-sourced assessments, independent verification of China's claims about proprietary innovation is absent. Watch for U.S.-China AI governance talks and any new U.S. chip export restrictions expected this quarter.
Coverage: 2 sources, 2 neutral
Sources (2): thehindu.com (neutral report), timesnownews.com (neutral report)
This story was synthesised by AI from the 2 sources linked above. Methodology and corrections.
Updated: this story now draws on 2 sources.
