Alibaba will ban employees from using Anthropic’s AI coding assistant Claude Code starting July 10, after determining that the software posed elevated security risks. An internal notice described Claude Code as having a “backdoor risk” and said it had been added to a list of high-risk software with security vulnerabilities. Reuters, citing an anonymous employee, reported that staff were told to switch to Alibaba’s own coding platform, Qoder.
The move followed reports from independent researchers and developers who said they found concealed code in Claude Code that could identify users connected to China. A Reddit user known as LegitMichel777 said the software checked whether a proxy was enabled and then altered the system prompt in ways that could transmit information such as whether the user was in China, whether a proxy connection involved a Chinese URL, and whether the user had an AI research institute in China. The user also alleged that the code was obfuscated within the binary. Similar findings were reportedly described by independent developer Thereallo.
Anthropic, through engineer Thariq Shihipar, said the feature was part of an experiment launched in March 2026 to curb account abuse by unauthorized resellers and guard against distillation, a practice in which one AI model is trained on another model’s outputs. Shihipar said stronger protections had since been introduced and that the company had already planned to remove the experiment, with a full rollback in a July 1 release. That explanation contrasts with critics’ characterization of the feature as hidden tracking or spyware.
The dispute comes amid broader tensions over access to Anthropic’s models in China. Anthropic already bars Chinese companies and foreign entities they own from using its models and has been trying to close loopholes that allowed access. The issue also follows Anthropic’s accusation that Alibaba’s AI research unit, Qwen Lab, carried out a distillation attack involving Claude. Alibaba and Anthropic had not publicly commented further on the latest dispute.