Anthropic accused Chinese firms of “industrial-scale distillation attacks” on its AI models.
Distillation involves training less capable models on more advanced ones’ output, and can be used illicitly to acquire powerful capabilities cheaply. The AI startup accused China’s DeepSeek, MiniMax, and Moonshot of generating “over 16 million exchanges with Claude through approximately 24,000 fraudulent accounts,” and said that the models thus trained would lack safeguards and could be used to develop weapons or carry out cybercrime.
OpenAI similarly accused DeepSeek of distillation attacks last year, after the Chinese firm shocked the world with the success of its cheap R1 model. Anthropic has been consistently in favor of export restrictions to slow China’s AI progress.