Anthropic Accuses Chinese Labs of Illicitly Distilling Claude Bot

Anthropic accused Chinese artificial intelligence firms DeepSeek, Moonshot, and MiniMax of conducting industrial-scale campaigns to illicitly extract capabilities from its Claude model.The labs allegedly used 24,000 fraudulent accounts to perform 16 million exchanges for model distillation, a process that trains cheaper models using data from advanced systems.These allegations follow similar claims from OpenAI as Anthropic warned that such practices bypass USA export controls and pose significant national security risks.