Alibaba has said that, from 10 July, employees will no longer be allowed to use Anthropic’s Claude Code. The Chinese technology group classed the coding tool as high risk and told staff to switch to Qoder, its own alternative. The decision reflects Alibaba’s security concerns and fits a wider pattern in which Chinese technology companies are tightening controls around foreign artificial intelligence tools.
Anthropic already bars Chinese companies, and foreign organisations they own, from using its models, and it is working to close loopholes that could let Chinese users access Claude systems. In a separate explanation on X, Anthropic’s Thariq Shihipar said that a version of Claude Code capable of secretly identifying Chinese users was part of an experiment begun in March. He said the aim was to block unauthorised account abuse and protect against model distillation, which is the process of training one AI model using the outputs of another. Anthropic has since taken stronger measures and plans to remove the feature, but Alibaba’s ban will still go ahead.
