Key facts
- Alibaba will ban employees from using Anthropic's Claude Code starting July 10.
- Anthropic already prohibits Chinese companies from using its models.
- Anthropic experimented with a tool to identify Chinese users, citing account abuse and distillation concerns.
- Alibaba has deemed Claude Code high-risk software.
- Employees are instructed to use Alibaba's Qoder tool.
China's Alibaba is reportedly set to ban its employees from using Anthropic's Claude Code programming tool, with the prohibition taking effect on July 10. This decision comes as Anthropic itself has measures in place to prevent Chinese companies and their foreign-owned entities from accessing its AI models.
Recent reports suggest Anthropic has been working to close loopholes that allowed Chinese users to access Claude. An experiment, described by Anthropic's Thariq Shihipar on X, involved a version of Claude Code that could identify Chinese users. Shihipar stated this was an experiment launched in March to prevent account abuse from unauthorized resellers and protect against model distillation, a process where AI models are trained on the outputs of other models. He added that stronger mitigations have since been implemented and the company had intended to remove the experimental feature.
Despite these developments, Alibaba has reportedly classified Claude Code as high-risk software. The company is now directing its employees to utilize its internal Qoder tool as an alternative.