China has warned users about alleged security vulnerabilities in Anthropic’s CloudCode AI coding tool, claiming it could transmit sensitive user information without authorization. The warning comes as Alibaba moves to ban the software for its employees over security concerns.
According to China’s National Vulnerability Database (NVDB), the alleged backdoor could allow the software to transmit sensitive user information, including location and identity data, to Anthropic’s servers without permission.
CloudCode is an AI coding agent that can generate computer code based on user instructions, debug software, and review code.
Anthropic has reportedly restricted access to its products in China and other countries it considers rivals, but it is still possible to use the tools through VPNs or third-party proxy services.
NVDB, which is affiliated with China’s Ministry of Industry and Information Technology, said on its website that it had recently discovered “security backdoor vulnerabilities” in the tool that could pose a “serious threat.”
According to a foreign news agency, Anthropic has not yet responded to a request for comment on the allegations.
NVDB has instructed organizations and users to immediately conduct a full investigation and either uninstall the software or update to the latest version of the software that removes the alleged backdoor.
On the other hand, Chinese technology company Alibaba has also decided to ban its employees from using “cloudcode” from July 10 due to security concerns.
It should be noted that Anthropic has previously accused Alibaba of copying its AI models.
