Alibaba to ban employees from using Claude Code: What happened?

Alibaba is reportedly set to ban employees from using Anthropic's Claude Code after security concerns and an ongoing dispute over AI model distillation. The company is instead asking staff to switch to its own AI coding platform, Qoder.

In Short

• Alibaba classifies Claude Code as "high-risk software," with the ban expected from July 10
• The move comes amid Anthropic's accusations that Alibaba attempted to distil its AI models
• Microsoft is also reducing Claude Code usage internally, but for cost-related reasons

Alibaba is reportedly asking its employees to stop using Anthropic's Claude Code for work, adding another chapter to the growing rivalry between Chinese and US artificial intelligence companies. The reported move comes as Anthropic and Alibaba remain locked in a dispute over AI model training practices, while concerns have also emerged over certain features in Claude Code that were designed to detect users attempting to access the service from restricted regions.

According to Reuters, Alibaba has internally classified Claude Code as "high-risk software" and plans to prohibit employees from using it at work. Employees are instead being directed to use Alibaba's own AI coding assistant, Qoder. The company has not publicly commented on the reported decision, and Anthropic has also not issued an official response regarding the reported workplace ban.

People familiar with the matter told Reuters that the decision follows recent scrutiny over Claude Code's built-in mechanisms that inspect user environments. Developers had claimed that the AI coding assistant checked information such as time zones and proxy-related details before sending prompts to Anthropic's servers. These findings sparked discussions online, particularly among users in China, where Anthropic officially restricts access to its AI services.

Responding to the concerns, Anthropic engineer Thariq Shihipar said on X that the feature was "an experiment we launched in March" and was intended to prevent account abuse by unauthorised resellers while also protecting the company's models from distillation efforts.

Morningstar reported that Alibaba has labelled Claude Code as "high-risk software" following these reports and plans to enforce the ban from July 10. Employees will reportedly be required to switch to Alibaba's in-house coding platform, Qoder.

Dispute between Anthropic and Alibaba

The reported workplace ban also comes amid rising tensions between the two companies. Last month, Anthropic accused Alibaba of attempting to extract the capabilities of its AI models through a process known as distillation, where a smaller AI model is trained using the outputs of a more advanced one.

In a letter sent to two US senators and reviewed by Reuters, Anthropic alleged that such efforts could help Chinese AI companies reach the capabilities of its advanced Mythos Preview models more quickly.

Earlier this year, Anthropic had also said it identified activity linked to accounts that it believed were associated with senior staff from several Chinese AI labs attempting to distil its models. The company has long restricted access to Claude in China, citing national security concerns. Despite those restrictions, Claude Code has remained popular among Chinese developers, many of whom reportedly use overseas servers to access the service.

According to Reuters, enforcing these restrictions on individual users remains difficult because internet traffic can be routed through servers located outside China. However, companies face greater legal and compliance obligations, making internal bans easier to implement.

Neither Alibaba nor Anthropic has publicly commented on the allegations surrounding model distillation or the reported workplace policy.

The reported move also comes as Chinese AI companies increasingly focus on domestic and open-source AI models such as DeepSeek, Alibaba's Qwen, Moonshot and Zhipu, reducing reliance on overseas AI systems.

Interestingly, Alibaba is not the only major technology company moving employees away from Claude Code. According to a recent report by The Verge, Microsoft is also scaling back the use of Claude Code across parts of the company, although for different reasons. The report said Microsoft plans to remove many Claude Code licences and encourage developers to use its own GitHub Copilot CLI instead. Unlike Alibaba's reported security-driven decision, Microsoft's transition is said to be aimed at consolidating its internal AI development tools while also reducing operating costs. Anthropic's AI models, however, will continue to be available through Microsoft's Copilot CLI platform.

- Ends