U.S. AI company Anthropic has accused three Chinese AI firms, DeepSeek, Moonshot AI, and MiniMax, of illegally using its Claude chatbot. The firms allegedly improved their own models through a process called “distillation.” According to Anthropic, around 24,000 fake accounts were created, carrying out over 16 million interactions, extracting advanced capabilities without authorization and in violation of access restrictions.
Anthropic warns this could undermine intellectual property rights and pose national security risks. The allegations come amid ongoing debates over AI chip export controls and intensifying U.S.–China competition in AI technology.