Home / Technology / China's AI 'Shortcut': U.S. Models Fuel Defense Tech
China's AI 'Shortcut': U.S. Models Fuel Defense Tech
31 Jul
Summary
- Chinese researchers use U.S. AI models to train domestic systems.
- Model distillation technique creates specialized, local AI.
- U.S. seeks to restrict China's access to advanced AI technology.
Chinese military researchers are reportedly using outputs from leading U.S. artificial intelligence models, such as those developed by OpenAI and Anthropic, to train their own domestic AI systems. This technique, known as "model distillation," involves using the outputs of powerful AI to train smaller, more specialized models that can be deployed locally.
This approach offers a way for China to advance its defense capabilities, including for surveillance, cyber warfare, and tactical decision-making, without the immense computing power required to build frontier AI models from scratch. Despite U.S. restrictions on advanced chips and technology, this method provides a shortcut for Chinese institutions to gain insights and close the AI gap.
Papers from institutions linked to the People's Liberation Army detail how these U.S. AI models are used to process sensitive information and develop specialized functions, such as image processing for drones and target recognition for maritime operations. This practice has become a point of contention ahead of U.S.-China AI governance talks.
