Alibaba orders staff to uninstall Anthropic's Claude Code over alleged location tracking

Alibaba Group has ordered employees to uninstall Anthropic’s Claude Code by July 10, adding the AI coding assistant to its internal “high-risk software” list and directing staff to switch to Alibaba’s own tool, Qoder.

The order follows claims that Claude Code carried code inspecting a user’s timezone and proxy settings to flag whether they were in China or connected to Chinese AI labs, then embedded markers in prompts sent to Anthropic’s servers. Developers and security researchers surfaced the findings this week on Reddit and GitHub.

Anthropic has not officially commented on the tracking claim, and the backdoor description comes from outside researchers, not a confirmed disclosure. An Anthropic researcher reportedly said the mechanism was added as an anti-abuse experiment aimed at account resellers and unauthorized model distillation, and had already been merged for removal in an upcoming update. That account, too, is unverified.

The dispute did not start here. In a late-June letter to US senators, Anthropic alleged that operatives linked to Alibaba’s Qwen AI lab ran roughly 25,000 fraudulent accounts, generating about 28.8 million interactions with Claude between April 22 and June 5 to distill a competing model. Separately, Anthropic is moving to close loopholes that let Chinese firms reach Claude through offshore entities, resellers and cloud platforms.

The uninstall deadline turns a technical fight over anti-abuse code into a corporate split. Whether other Chinese firms follow Alibaba, and whether Anthropic ships the promised removal, will show how far the access standoff spreads.