Alibaba Bans Employees From Using Anthropic's Claude Code Over Data Leaks

Alibaba Issues Internal Ban

Alibaba Group has prohibited its employees from using Anthropic's newly released developer tool, Claude Code. The Chinese e-commerce and cloud giant classified the command-line interface as high-risk software due to potential security vulnerabilities. This decision reflects growing corporate anxiety over where proprietary code goes once fed into external artificial intelligence models.

The restriction was communicated to engineering teams via internal channels. Employees are no longer permitted to download, install, or run the tool on company-issued devices or within corporate networks.

Security and IP Concerns

Anthropic launched Claude Code as an agentic tool designed to write, debug, and explain code directly in the terminal. While developers praise its speed, the tool requires deep access to local file directories and network permissions to function. This level of access triggered alarms within Alibaba's cybersecurity division.

Several factors drove the Chinese tech giant's swift restriction:

• Data exfiltration risks: Code repositories often contain proprietary algorithms and internal API keys.
• Regulatory compliance: Chinese firms face strict domestic data residency laws that conflict with sending data to US-hosted AI servers.
• Alternative internal tools: Alibaba operates its own proprietary large language models, Tongyi Qianwen, and prefers internal developer adoption.

A Growing Corporate Trend

Alibaba is not alone in restricting external AI tools for software development. Major financial institutions, healthcare providers, and technology firms globally have implemented similar bans on third-party coding assistants. Apple, Samsung, and Amazon previously restricted internal use of OpenAI's ChatGPT and GitHub Copilot for similar intellectual property reasons.

These restrictions highlight a growing tension between developer productivity and corporate security protocols. While agentic AI tools speed up software delivery, the potential cost of a proprietary code leak remains too high for major enterprises.

Industry analysts expect more tech conglomerates to block external command-line AI tools in favor of self-hosted, private alternatives.