Chinese tech giant Alibaba has barred its employees from using Anthropic's Claude Code, the company's AI-powered coding assistant, amid allegations that the tool contains a hidden "China-detection" backdoor.
According to a report surfaced via Bing News (published on MSN), the ban followed the discovery of an alleged hidden mechanism designed to detect when the software was being used in China. The report does not detail how the alleged backdoor was found or verified.
The move appears to be part of a widening rift between the two companies. The MSN report notes that the ban lands roughly three weeks after Anthropic publicly accused Alibaba's Qwen lab of running what Anthropic called the largest known "distillation attack" on Claude — a technique in which one AI model is used to train or copy the capabilities of another.
Separately, VOI.id (carried via Google News) reports that Alibaba employees have not only been blocked from Claude Code but have been asked to switch to Qoder, an alternative coding tool.
Taken together, the two sources describe a tit-for-tat escalation: Anthropic accusing Alibaba's AI lab of improperly copying its model, followed by Alibaba pulling Anthropic's product from its workforce and steering staff toward a homegrown substitute. The specific claims on each side — the alleged distillation attack and the alleged backdoor — have not been independently confirmed in these sources.
Why it matters: The dispute signals how quickly trust is breaking down between American AI developers and major Chinese tech firms, pushing both sides toward separate, incompatible software ecosystems.