On Friday, Anthropic found itself in hot water after a security researcher uncovered a hidden tracker in its Claude Code, raising concerns over user privacy. The tracker, which monitored Chinese users, was revealed by the developer known as “Thereallo,” who criticized it as a “serious breach of user trust.” Anthropic confirmed the code’s existence, stating it was part of an experiment aimed at preventing account abuse.
Background on the Claude Tracker Incident
The tracker was added in March as a measure to combat unauthorized reselling of access to Claude Code, which has been sold for as little as $12, compared to the official subscription rate of up to $100 per month. Anthropic engineer Thariq Shihipar explained that the tracker was intended to protect against “distillation attacks,” a method by which other AI firms could replicate their models.
Despite this explanation, privacy advocates expressed outrage, arguing that Anthropic's actions contradict its stated commitment to user privacy. The tracker was designed to collect minimal data, such as users’ time zones and proxy information, but its stealthy implementation raised eyebrows among users and experts alike.
Concerns Over User Surveillance
Anthropic's decision to monitor users has sparked a debate about ethical boundaries in AI development. The firm previously opposed U.S. government surveillance, particularly against its users. The irony is not lost on observers, as Anthropic now faces accusations of engaging in similar practices that it once condemned. The company has even filed a lawsuit against the White House over its refusal to allow government access to Claude.
