China Flags ‘Security Backdoor’ in Anthropic AI Coding Tool, Claude Code

Beijing – A Chinese industry regulator issued a warning on Wednesday regarding a “security backdoor” embedded in versions of Anthropic’s AI coding tool, Claude Code. The alleged backdoor could allow the software to transmit sensitive information, including users’ locations and identity-related identifiers, back to Anthropic’s servers without user consent, according to China’s National Vulnerability Database (NVDB).

Claude Code is an AI coding agent that can generate computer code, debug software, and review code based on user prompts. Anthropic blocks access for users in China and other nations it deems adversarial but the tool remains accessible through VPN or third-party proxy services.

The NVDB, affiliated with China’s Ministry of Industry and Information Technology, stated on its website that it had detected “severe security backdoor risks” in Claude Code. Anthropic has not responded to AFP requests for comment on these allegations.

The NVDB advised institutions and users to conduct a comprehensive check and uninstall or upgrade to the latest secure version. It also urged organizations to strengthen network traffic monitoring to prevent unauthorized data leakage.

Chinese tech giant Alibaba informed employees last week that it would ban Claude Code starting July 10 due to security concerns, according to people familiar with the matter. Anthropic has previously accused Alibaba of reverse-engineering its AI models for “distillation.”

Claude Code engineer Thariq Shihipar responded in an X post, stating that reports about tracking Chinese users’ data were an experiment launched in March to prevent account abuse from unauthorized resellers and protect against distillation. The team has since implemented stronger mitigations and the feature will be fully rolled back in tomorrow’s release.