China Warns of 'Security Backdoor' in Anthropic AI Coding Tool

China has issued a warning regarding a potential "security backdoor" within Claude Code, an artificial intelligence coding tool developed by the US company Anthropic. The National Vulnerability Database (NVDB), a cybersecurity platform affiliated with China's Ministry of Industry and Information Technology, alerted users on Wednesday.

According to the NVDB, the alleged vulnerability could enable the software to transmit sensitive user data, such as locations and identity-related identifiers, back to Anthropic's servers. This transmission could occur without the explicit consent of the users, raising significant privacy and security concerns.

Claude Code is designed to assist developers by generating computer code, debugging software, and reviewing code based on user prompts. While Anthropic officially restricts access to its products for users and companies in China and other nations it considers adversarial, the NVDB noted that the tool can still be accessed within China. This is often achieved through the use of Virtual Private Networks (VPNs) or third-party proxy services, circumventing the company's geographical restrictions.

The NVDB's advisory highlights ongoing tensions and security considerations surrounding the use of advanced AI technologies developed in the West within China. The potential for data exfiltration, even through indirect access methods, is a key concern for Chinese cybersecurity authorities.