Anthropic is now rolling back a hidden tracking feature in its AI coding assistant Claude Code after the developers accused the company of secretly identifying users who were in Chine or connected with Chinese labs. Anthropic said that the system was introduced to tackle account abuse and prevent its AI models from being copied, but the firm has been facing enormous criticism. As per The Information, Anthropic has stated that the feature was an experiment built to curb account abuse and prevent rival AI firms from exactly copying its models.
Furthermore, the company has constantly said that multiple Chinese AI firms like Moonshot AI, Alibaba, Minimax, and DeepSeek have attempted large scale distillation attacks on Claude. For those unaware, Claude is officially unavailable in China, and Anthropic has been introducing stricter rules against users bypassing its geographic restrictions. Nonetheless, a lot of developers and security researchers debated that the tracking system in Claude was integrated without proper disclosure.
The researchers said that Claude Code has been quietly collecting their location data and also appeared to identify users based on where they were connecting from or which organisations they were linked to, raising concerns related to user privacy. As per the reports, the feature tracked the time zones and proxy settings of the users to identify if they were connecting from China or Chinese AI networks.
In response, Anthropic has not denied the existence of such feature nor they have agreed that it exists. One of the Claude Code Engineers said that this is an experiment that started in March. He wrote on X, ‘This is an experiment we launched in March that was meant to prevent account abuse from unauthorised resellers and protect against distillation.’ He also added that the firm has already developed stronger protections and had been planning to remove the hidden market system super soon.
