Alibaba will bar employees from using Anthropic's AI tools for work from 10 July, citing security concerns, according to an internal notice reported by the South China Morning Post and confirmed by CNBC.
The company has classified Claude Code, Anthropic's AI coding assistant, as high-risk software. Staff have reportedly been told to remove Anthropic's models from their work environment and adopt Qoder, Alibaba's own coding platform, instead.
The ban follows two escalations. In a letter to US senators dated 10 June, Anthropic accused operators linked to Alibaba's Qwen AI lab of running a large-scale 'distillation' campaign, generating 28.8 million exchanges with Claude through roughly 25,000 fraudulent accounts to train a rival model. Alibaba has denied wrongdoing. Days before the ban, developers also documented undisclosed detection features inside Claude Code that could inspect signals such as timezone and proxy information; an Anthropic engineer acknowledged the mechanism as an anti-abuse experiment and said it has been removed.
The dispute signals a hardening split between US AI firms and Chinese technology companies, with employees' everyday tools now caught in the middle. Wocult's full analysis of the fight, and what it means for Indian companies, is here.
