China warns of "security backdoor" in Anthropic AI coding tool [July 8, 2026]
OttieneChina's vulnerability database flags Claude Code as a backdoor risk
China's National Vulnerability Database (NVDB), which sits under the Ministry of Industry and Information Technology, said Wednesday that versions of Claude Code carry "backdoor risks posing a serious threat." Per AFP, the regulator claims the tool can send geolocation data and user identifiers to Anthropic's servers without consent. NVDB told organizations to run a full audit immediately and either remove Claude Code or update to the latest secure version.
Claude Code engineer Tarek Shihipar responded on X that the behavior comes from an experiment launched in March to guard against unauthorized resellers and distillation, and that it "should be fully rolled back in tomorrow's release."
Alibaba has warned staff that Claude Code is off limits as of July 10, according to people familiar with the situation.
Up next
River AI raises $1.1 billion to build "personal AI"
River AI raised $1.1B to build personal AI that each user owns and tunes themselves, in a round co-led by General Catalyst and AMP PBC with strategic checks from NVIDIA and AMD Ventures. It sells an API platform for fine-tuning on open models.
Related stories
Comments
No comments yet. Be the first.
Join the conversation
Sign in with Google to leave a comment. Your name and avatar come from your Google profile, and the comment appears after moderation.