Anthropic accused three prominent Chinese AI companies—DeepSeek, Moonshot AI, and MiniMax—of creating over 24,000 fraudulent accounts to exploit its Claude chatbot, sending more than 16 million prompts to refine their own models, according to a blog post from February 23. This practice, known as “distillation,” allows these firms to replicate advanced AI capabilities without the substantial investments required for independent development. Anthropic warned that such activities threaten the competitive edge of U.S. AI companies, as many Chinese models do not charge a subscription fee, potentially undermining U.S. market monetization efforts. The allegations also coincide with concerns over U.S. chip export policies that have recently eased restrictions, which Anthropic argues could facilitate further illicit AI advancements in China.
Anthropic accuses DeepSeek, Moonshot, MiniMax of fraudulent access to Claude
