Anthropic quickly removed a hidden tracker from its Claude Code tool after a security researcher exposed it as a serious breach of user trust. The tracker used prompt steganography to secretly monitor Chinese users' timezone, proxy, and potential connections to AI labs that Anthropic has accused of distillation attacks.
An Anthropic engineer confirmed the tracking was a March experiment aimed at preventing account abuse and unauthorized reselling. The incident has drawn criticism for contradicting Anthropic's anti-surveillance stance.
Comments