Beijing has raised concerns about a “security backdoor” in Anthropic’s Claude Code, an AI tool used for generating and debugging computer code. The Chinese National Vulnerability Database has identified significant security risks, urging users to check for vulnerabilities and update to secure versions. Anthropic argues that the mechanism serves as an anti-abuse measure, checking device timezones and routing to prevent fraud. This controversy follows past tensions with Alibaba, accused of reverse-engineering Anthropic’s AI models. Anthropic plans to remove the contentious feature in an upcoming update.
Is there a ‘security backdoor’ in Anthropic’s AI coding tool?
