Alibaba prohibits employees from using Anthropic AI following ‘distillation’ dispute

Alibaba is removing Anthropic’s AI tools from its workplace, becoming one of the latest tech companies to tighten control over the software employees can use.

The Chinese technology giant cited security concerns and classified Claude Code as high-risk software.

The decision is the latest sign of the deepening AI rivalry between the US and China.

In June, the US startup accused Alibaba of trying to extract capabilities from its AI models through a technique known as “distillation,” in which one model learns from another.

In a letter to the Senate Banking Committee, Anthropic alleged that Alibaba created 25,000 fraudulent accounts that generated roughly 28.8 million exchanges with Claude to help develop its own AI products.

Anthropic described the incident as the largest known attempt of its kind and said its terms prohibit companies in China and other countries it considers adversarial from accessing its models.

Reuters reported that Chinese authorities met with Alibaba, ByteDance, Z.ai, and other leading AI firms to discuss curbing overseas access to their latest and future AI models.

Officials at the meetings were reportedly “deeply” concerned that Anthropic’s Mythos model could exploit security vulnerabilities and be weaponized by the US against China. US lawmakers have echoed similar concerns about advanced American AI technology falling into the hands of rivals such as China and Russia.

Meanwhile, Anthropic has tightened enforcement of regional restrictions by blocking attempts to access its AI services through overseas offices or virtual private networks (VPNs).