Anthropic Accuses Chinese AI Labs of Systematic Claude Extraction

Anthropic has publicly escalated its accusations against Chinese AI rivals, detailing what it describes as systematic, coordinated campaigns to extract the intellectual property of its Claude models. The company specifically named DeepSeek, Moonshot AI, and MiniMax, alleging they used tens of thousands of fake accounts to interact with Claude's API. This activity, termed 'model extraction,' is akin to reverse-engineering a product through intensive, automated probing. By submitting millions of prompts and analyzing the outputs, a competitor could theoretically train their own model to replicate key capabilities of Claude at a lower cost. Anthropic's public statement marks a significant hardening of its position, transforming a private enforcement issue into a public allegation of unfair competitive practices. The move occurs against a backdrop of heightened U.S.-China technological competition and ongoing policy debates about restricting China's access to AI chips and tools. By going