China’s CNVDB, a state-run cybersecurity vulnerability database, said on July 8 that Claude Code versions released from April to June could send user location and identity-related data to remote servers without consent. It urged users to delete the software or update to a newer version.

Alibaba told employees it would ban workplace use of Claude Code from July 10. Anthropic said Chinese companies such as Alibaba were not authorized to access Claude services.

The notice followed online claims that Anthropic added code to detect users in China. A Claude Code engineer said on X that the code was part of a March experiment to curb account abuse by unauthorized resellers and model distillation, and was being rolled back.

Anthropic said Claude Code includes a customer-controlled monitoring feature that is separate from the undisclosed data transmission alleged by Chinese authorities.

Claude is not directly offered in China, and access has relied on virtual private networks or proxy services.

🔗 Source: Chosun Daily

Recent Anthropic developments