Chinese technology giant Alibaba Group Holding has banned employees from using Anthropic’s AI coding assistant, Claude Code, for work, citing potential security risks. The company said the tool has been classified as high-risk software following an internal security review, with the ban set to take effect across Alibaba offices from July 10.
According to an internal notice, Alibaba said Claude Code was found to pose “back-door risks” and has therefore been added to the company’s list of software with security vulnerabilities.
The decision comes after reports claimed that Anthropic had embedded hidden code in Claude Code to identify users based in China or those connected to Chinese AI laboratories. Security researchers raised the issue through posts on Reddit and GitHub, alleging that the tool collected signals such as Chinese time-zone settings and proxy server usage before sending the information back to Anthropic through concealed messages.
Responding to the controversy, Anthropic engineer Thariq Shihipar said the feature was introduced as an experimental measure in March. According to him, it was designed to prevent misuse by unauthorised resellers and to protect Anthropic’s AI models from “distillation”—a process in which outputs from a large AI model are used to train smaller and less expensive models.
Anthropic has previously accused several Chinese AI companies, including Alibaba, of attempting to replicate the capabilities of advanced US AI models through distillation techniques.
Shihipar said the tracking mechanism would be removed in an upcoming update of Claude Code, adding that the company had already planned to discontinue the feature after developing stronger security safeguards.
The controversy comes at a time when the United States is tightening restrictions on advanced AI technologies. Last month, Washington imposed export controls on Anthropic’s Fable 5 AI model after security concerns were reportedly identified. The company temporarily disabled the model before restoring access after the restrictions were eased.
Alibaba’s move also highlights the growing effort by Chinese technology firms to reduce dependence on American AI software. The company has reportedly recommended that employees switch to Qoder, Alibaba’s own AI coding assistant, instead of Claude Code.
Although Anthropic’s AI products are not officially available in China, many developers continue to access them through third-party platforms and other workarounds because of their strong coding and reasoning capabilities.
The latest development underscores the increasing geopolitical and cybersecurity tensions surrounding artificial intelligence, as both the US and China seek greater control over advanced AI technologies and their global use.
Recent Random Post:
