Alibaba tells employees to stop using Anthropic’s Claude Code

Alibaba Group Holding has barred employees from using Anthropic’s Claude Code for work, citing security risks linked to the US artificial intelligence company’s earlier use of hidden code to identify users in China.

The Chinese technology giant said Claude Code had been added to its list of high-risk software following an internal review. The ban will take effect across Alibaba offices from July 10.

“As Claude Code was recently discovered to carry back-door risks, after comprehensive evaluation, Claude Code has now been added to a list of high-risk software with security vulnerabilities,” Alibaba said in the notice.

The move follows allegations that Anthropic embedded code in its AI coding tool to detect whether users were based in China or linked to Chinese AI laboratories.

Security researchers flagged the practice earlier this week through posts on Reddit and GitHub. The code allegedly collected signals such as Chinese time-zone settings and the use of proxy servers, and passed the information back to Anthropic through hidden messages, according to cybersecurity publication International Cyber Digest.

Anthropic engineer Thariq Shihipar responded to the criticism on X, describing the feature as an “experiment” introduced in March.

He said it was intended to prevent misuse by unauthorised resellers and protect Anthropic’s models from distillation. Distillation is a technique in which the outputs of a larger AI model are used to train a smaller and cheaper model.

Earlier, Anthropic accused Chinese AI companies, including Alibaba, of using distillation to replicate the capabilities of advanced US models.

Shihipar said the tracking mechanism would be removed as part of a broader re-release of Claude Fable 5. He added that Anthropic had been planning to discontinue it after developing stronger safeguards.

The controversy comes amid growing restrictions on the use of advanced US AI systems outside the country.

Washington last month imposed export controls on Fable 5, the public version of Anthropic’s more powerful Mythos model, after security vulnerabilities were reportedly identified by Amazon researchers. Anthropic temporarily disabled the model for all users before restoring access after the restrictions were lifted on Thursday.

The company said it would work with the US government on improving the security of frontier AI systems.

Alibaba’s decision also reflects a broader push among Chinese technology companies to reduce their reliance on US software and AI models.

The company has recommended that employees use Qoder, its own AI coding agent, as an alternative to Claude Code, according to people familiar with the matter.

Anthropic’s products are not officially available in China. However, Chinese developers have continued to access them through third-party services and other workarounds, attracted by their coding and reasoning capabilities.