Anthropic outed for Claude tracker that secretly monitored Chinese users

Anthropic removed a hidden tracker from its Claude Code application after a security researcher discovered the code and called it a “serious breach of user trust” [1]. The company acknowledged the tracker had been active since March as an “experiment” to curb account abuse and distillation [1]. The researcher, a web developer known as “Thereallo,” found Anthropic was using “prompt steganography” to conceal code that flagged users in China [1]. The mechanism relied on shorthand markers to quietly collect timezone and proxy data, and to detect potential connections to Chinese AI labs that Anthropic has accused of distillation attacks [1]. Thereallo described the practice as a “serious breach of user trust” [1]. Anthropic engineer Thariq Shihipar confirmed on X that the tracker was added to Claude Code as an “experiment” in March [1]. Shihipar said the code “was meant to prevent account abuse from unauthorized resellers and protect against distillation” [1]. The Washington Post reported that unauthorized retailers have sold access to free models for $1 a month, and pro subscriptions that normally cost $100 monthly for “as little as $12” [1]. Shihipar added that the company had “actually been meaning to take this down for a while” because engineers have “landed stronger mitigations since then” [1]. Anthropic, founded in 2021 by former OpenAI members, has positioned itself with a focus on AI safety and developed the Claude series of large language models [6]. The company was valued at an estimated $965 billion in May 2026, making it the most valuable pure-play AI firm in the world [6]. Claude is trained using “constitutional AI,” a technique designed to improve ethical and legal compliance [7]. The revelation arrives during a period of tension between Anthropic and the U.S. government. Since January 2026, the Department of Defense has been in conflict with the company over the use of its products for military purposes and mass domestic surveillance [8]. U.S. federal agencies began phasing out Claude after Anthropic refused to remove contractual prohibitions on using the technology for mass domestic surveillance and fully autonomous weapons [7]. The Department of Defense subsequently designated the company a “supply chain risk” and barred private military contractors from doing business with the firm, though a federal judge issued a temporary injunction against that designation on March 26, 2026 [7]. Privacy advocates warned that the hidden tracker shows Anthropic is willing to cross lines to surveil users, a stance that appears at odds with the company’s public refusal to allow the U.S. government to use Claude for domestic surveillance [1]. Anthropic has sued the White House over the clash [1].