Alibaba Reportedly Bars Employees From Using Anthropic’s Claude Code

An unconfirmed internal ban would reflect continued corporate caution over alleged backdoor risks and licensing limits even after Anthropic restored its suspended models and added new safeguards.

Overview

Two news outlets report that Alibaba has decided to block employee use of Claude Code inside its workplace systems, though the company has not issued an official statement.
The reported restriction is said to target alleged embedded backdoors in the coding assistant, a claim sourced to an unnamed person familiar with the decision and not independently verified.
This development follows recent moves by large firms, including banks in Hong Kong, to limit or remove Anthropic models over licensing and regional-use concerns.
Anthropic restored access to its Fable 5 and Mythos 5 models after U.S. export limits were lifted and said it added classifiers to block more cybersecurity tasks and will expand cooperation with U.S. authorities on testing and misuse tracking.
If confirmed, the Alibaba step could deepen enterprise reluctance to deploy third-party code assistants, affect developer workflows inside large firms, and prompt more formal vendor audits or regulatory scrutiny.