Alibaba Blocks Anthropic's Claude Code Amid Security Allegations

The controversy erupted after a Reddit post alleged that a version of the software released in April included mechanisms to track users within Chinese borders. Thariq Shihipar, a developer at Anthropic, clarified on X that the code was part of a March experiment intended to curb unauthorized account reselling and protect against model distillation—a process where developers train their systems using the outputs of another company's AI.

Anthropic has previously accused Chinese laboratories, including those affiliated with Alibaba, of illicitly distilling its models. While Anthropic restricts direct access to Claude in China citing national security concerns, the tool remains widely utilized by the country's engineering and research communities. By classifying the software as high-risk, Alibaba aims to insulate its internal infrastructure from potential surveillance and the ongoing intellectual property disputes between the U.S. AI firm and Chinese developers.