Post by : Badri Ariffin

Alibaba, a major player in China's tech industry, has mandated that employees cease utilizing Anthropic's AI coding tool, Claude Code, for work-related tasks, as per sources close to the company's decision.

Employees are now directed to transition to Alibaba's proprietary AI coding solution, Qoder. The source requesting anonymity cited confidentiality concerns.

Neither Alibaba nor Anthropic provided immediate feedback to inquiries from Reuters. Alibaba has yet to address the allegations raised by Anthropic publicly.

Questions Surround User Environment Insights

This move comes amidst heightened scrutiny of Claude Code, particularly following reports that developers found features allowing the tool to analyze user environments, such as their timezone and proxy settings. Concerns were raised that subtle markers were inserted into communications sent back to Anthropic.

On Tuesday, an Anthropic team member disclosed on X that these features were part of a March experiment intended to deter unauthorized usage by resellers and protect Anthropic's AI from quality degradation.

Allegations of Model Distillation

This development intensifies the ongoing conflict between Alibaba and Anthropic concerning AI technologies.

Last month, Anthropic accused Alibaba of engaging in "distillation"—alleging that the latter aimed to develop a less sophisticated AI model based on outputs from Claude AI.

A letter, reviewed by Reuters and sent to U.S. senators, indicated that this practice could enable China to accelerate its advancements in AI technology, specifically regarding its Mythos Preview model.

Alibaba has yet to publicly counter these accusations.

Popularity of Claude Code in China

Despite restrictions, Claude Code, designed primarily for developers, has found a significant user base in China. Although Anthropic limits access for Chinese entities, many programmers have circumvented these barriers.

The insider mentioned that enforcing these restrictions is a challenge since users can operate servers based in the U.S. and manipulate internet traffic locations.

However, businesses tend to exercise caution due to potential legal and compliance hurdles.

Shift Toward Domestic AI Solutions

As U.S. firms combat unauthorized use and model distillation, Chinese AI and cloud computing companies are increasingly embracing local and open-source solution alternatives.

Notable examples include Alibaba's Qwen, Moonshot, and Zhipu.

Simultaneously, Chinese AI developments are gradually making headway into the U.S. market, raising alarms among certain sector experts.

Initial Reports by Chinese Media

The ban on Claude Code use by Alibaba employees was initially disclosed by media outlets in China.

#Global News #Tech News #AI Technology #World News