Alibaba is set to ban employees from using Anthropic's Claude Code starting July 10, citing an alleged backdoor discovered by a researcher who reverse-engineered the tool. The alleged mechanism, reportedly active since version 2.1.91 (April 2), quietly checked whether a user's proxy or timezone matched a hidden list of Chinese corporate networks and AI labs — including Alibaba, Baidu, ByteDance, and Moonshot AI — and encoded the detection in the system prompt rather than sending overt telemetry. An Anthropic team member said the feature was intended to combat account reselling and model distillation, and that it would be removed. The ban comes amid an escalating dispute between Anthropic and Alibaba's Qwen lab, with Anthropic having accused Qwen-linked operators of running nearly 25,000 fraudulent accounts to extract Claude's capabilities in a large-scale distillation campaign. No independent security audit has been published, and both companies have yet to issue formal public statements.
3.5K Impressions1 Comment