Alibaba Bans Claude Code After Detecting Hidden Tracking Code

Alibaba Group Holding will ban its employees from using Anthropic's Claude Code AI assistant starting July 10, 2026. The decision follows the discovery of hidden tracking code in the April release (version 2.1.91) that used steganography, time zones, and proxy URLs to identify users based in China or affiliated with Chinese AI labs. Alibaba classified the tool as high-risk software with security vulnerabilities, directing staff to uninstall all Anthropic products and transition to its proprietary AI coding platform, Qoder.

Anthropic defended the feature as a March experiment intended to prevent account abuse from unauthorized resellers and protect against distillation, where rivals train models on another's output. The company stated the code was removed on July 1.

The dispute escalates after Anthropic accused Alibaba's Qwen AI lab of conducting a massive-scale distillation campaign between April and June, allegedly using 25,000 fraudulent accounts to perform 29 million exchanges. Anthropic also identified other Chinese firms, such as ByteDance and Ant Group, using VPNs, reimbursement schemes, and Microsoft Azure cloud services to bypass regional restrictions. While Alibaba denied the distillation allegations, the conflict highlights intensifying AI rivalry and security tensions between the United States and China.