Anthropic has been accused of embedding hidden code in its Claude Code platform to monitor Chinese users, according to a report by International Cyber Digest. The code, described as resembling spyware, allegedly collected data such as time zone, proxy, and routing metadata without user detection. This has been labeled a "serious breach of trust" due to its covert nature and the permissions involved. In response, Anthropic employee Thariq stated that the code was part of an experiment initiated in March to prevent unauthorized account use and model distillation. He assured that stronger anti-abuse measures are now in place and that the code responsible for the data collection has been rolled back, with complete removal expected in the next software update.