Anthropic Claims Alibaba Ran 'Brazen' Campaign to Access Its Claude AI Model

What We Know

Anthropic, the developer of the Claude family of AI models, has publicly accused Alibaba of mounting a deliberate campaign to illicitly access and extract Claude’s capabilities. Multiple outlets report Anthropic characterizing the effort as large-scale and “brazen,” and saying it amounted to a distillation or cloning attempt on Claude. According to reporting that cites Anthropic’s claims, the campaign allegedly involved Alibaba’s Qwen lab using about 25,000 fake accounts to make nearly 29 million exchanges with Claude — described as the largest such distillation effort yet. Coverage frames the activity as targeted model-extraction behavior rather than ordinary usage, and at least one outlet conveys Anthropic’s position that Alibaba should be held accountable for the operation. The dispute is drawing attention across business and tech press, with news organizations repeating Anthropic’s allegation that Alibaba “illicitly extracted” model capabilities. Reporting connects the episode to broader questions about model theft, distillation techniques (using many queries to reproduce behavior), and competitive risk in the commercial AI market.