Alibaba bans Claude Code AI coding tool over security concerns

Alibaba, the Chinese e‑commerce and cloud giant, will prohibit its employees from using Anthropic’s Claude Code coding assistant beginning 10 July. The company classified the tool as high‑risk software after internal reviews raised concerns that the tool contained mechanisms to detect China‑linked users – such as checking time‑zone and proxy data – which could act as backdoors or enable data leakage. Alibaba has ordered staff to switch to its own AI‑powered coding platform, Qoder.

The ban follows Anthropic’s recent tightening of access for Chinese entities and accusations that Chinese firms, including Alibaba’s AI lab, were using Claude outputs to train competing models via large‑scale “model distillation” with fraudulent accounts. Anthropic said the detection feature was an anti‑abuse experiment launched in March and intended to be removed. The move underscores growing corporate scrutiny of third‑party AI tools, heightened US‑China AI rivalry, and a push for domestic alternatives to foreign AI services.