The ban is the latest step in an escalating conflict between the two companies. Tensions rose in June when Anthropic accused Alibaba of conducting a so-called distillation attack. According to the US-based AI company, Alibaba illegally used data from their advanced models to train its own AI model.
Shortly after these accusations, it was discovered that Claude Code also contained hidden functions that reviewed users' time zones and proxy information. The purpose was to identify users with ties to China.
Regarding the tracking code, Anthropic employee Thariq Shihipar confirmed on platform X that the function was an experiment launched in March. The goal of the code was to prevent account abuse and protect the models against illegal distillation.
The team has implemented stronger measures since then and we have actually been planning to take this down for a while, Shihipar said in a statement.
Neither Alibaba nor Anthropic has yet officially commented on the reports of the internal ban.