Alibaba Group Holding Ltd. (Alibaba (BABA)) has decided to bar its employees from using AI tools developed by Anthropic, citing potential security risks. The ban takes effect on July 10.
The Chinese tech giant has labeled Anthropic's Claude Code as high-risk software. Employees have been directed to uninstall all Anthropic models and agent products and transition to Alibaba's in-house AI assistant, Qoder, according to CNBC on Monday.
Alibaba and Anthropic did not immediately respond to requests for comment.
China Access Under Scrutiny
This move follows Anthropic's allegations that Alibaba had tried to "brazenly" and "illicitly" distill its AI capabilities. In June, Anthropic sent a letter to the U.S. Senate Committee on Banking, Housing, and Urban Affairs, accusing Alibaba of launching "the largest known distillation attack" on its systems. Anthropic's terms of service prohibit Chinese companies and other "adversarial nations" from using its models. "Distillation" refers to training a less advanced AI model using outputs from a more capable system.
Meanwhile, according to the Financial Times on Friday, Anthropic is closing loopholes that allowed Chinese companies to access Claude through third countries after online backlash in China over hidden code designed to detect users based in the country. Some firms had accessed Claude via overseas corporate accounts or reimbursed employees for personal subscriptions used with VPNs.
The report says Ant Group gave employees access to Anthropic's Claude through corporate accounts linked to its Singapore entity, while TikTok parent ByteDance does not officially provide Claude access but reimburses engineers for personal subscriptions accessed via VPNs to help them learn and experiment with a broader range of AI tools.
