The US government accused Chinese artificial intelligence firms of unlawfully appropriating proprietary technologies from American AI developers through a process termed “distillation,” heightening tensions ahead of a critical bilateral summit between the world’s two largest economies later this month.
According to a joint statement from US law enforcement and intelligence officials, Chinese companies utilized distillation—a method involving training AI models on outputs from larger, more complex systems—to replicate and commercialize American AI innovations more efficiently.
The investigation named six entities, including DeepSeek, Moonshot AI, and Alibaba Group, for allegedly leveraging modified versions of US-developed AI models to accelerate their own product development. These actions were described as occurring in coordination with or under the awareness of Chinese government authorities.
Targeted American AI systems included those from Anthropic, OpenAI, Alphabet’s Google, and private aerospace firm SpaceX. US officials characterized the activity as a systematic effort to undermine intellectual property rights while circumventing expensive research and development investments.
“China-based AI enterprises are conducting large-scale, targeted distillation operations that strategically exploit Western innovations,” the officials stated, underscoring concerns about the erosion of competitive advantage in the global AI landscape.