China warns of "security backdoor" in Anthropic AI coding tool
Summary
China's cybersecurity agency warned that Anthropic's AI coding tool, Claude Code, may have a hidden backdoor that could send users' private information to the company without permission. The agency advised users to check their software and update or remove unsafe versions. Anthropic has said it is removing the feature that caused the concern.Key Facts
- China's National Vulnerability Database (NVDB) claimed Claude Code has a security backdoor risk.
- The backdoor could send sensitive data like user locations and identity information to Anthropic servers.
- Claude Code is an AI tool that helps write, debug, and review computer code.
- Anthropic restricts access to its products in China but users can still connect via VPN or proxies.
- NVDB advised users to uninstall or update Claude Code to a safe version immediately.
- Chinese company Alibaba banned Claude Code for its employees starting July 10 due to security fears.
- Anthropic engineer said the backdoor was an experiment to prevent abuse and would be removed soon.
- Anthropic accused Alibaba of copying its AI models using a method called "distillation."
Read the Full Article
This is a fact-based summary from The Actual News. Click below to read the complete story directly from the original source.
