Alibaba Bans Employees From Using Claude Code Over Hidden Tracking Concerns

Alibaba has told employees they have until July 10 to uninstall Claude Code and all Anthropic products from company computers. The Chinese company placed the tool on its internal list of high-risk software, claiming it detected a mechanism that covertly identified users linked to China. In its place, employees are being told to switch to Qoder, Alibaba's own coding tool.

What was detected

According to reports from Reuters and the South China Morning Post, Claude Code checked elements of the user's environment, including time zone and proxy settings, then inserted subtle markers into requests sent to Anthropic's servers. To Alibaba, this looked like a hidden backdoor for tracking Chinese accounts, so the company classified the tool as a security threat.

Anthropic does not deny the mechanism exists, but offers a different explanation. Company employee Thariq Shihipar wrote publicly that it was an experiment launched in March meant to prevent account abuse by unauthorized resellers and to guard against model distillation. He added that the team has since rolled out stronger safeguards and had been planning to retire this particular mechanism for some time.

The distillation dispute

The ban didn't come out of nowhere. In a letter to US Senators Tim Scott and Elizabeth Warren dated June 10, Anthropic accused entities linked to Alibaba and its Qwen lab of running the largest known distillation campaign against Claude. According to Anthropic, close to 25,000 fake accounts generated more than 28.8 million interactions with Claude models between April 22 and June 5, targeting in particular advanced coding and multi-step agentic reasoning.

Distillation involves flooding a model with carefully constructed queries, collecting its responses, and training a cheaper rival system on them to approximate the original's capabilities. Anthropic itself admits it has no hard proof, only an API traffic pattern and inference about who was behind the accounts, while Alibaba has not publicly addressed the distillation accusation.

Chain reaction

After the accusation, Anthropic began mass-blocking accounts linked to China, further straining relations between the two companies. Alibaba's decision to ban Claude Code is a response from the other side, driven partly by security concerns and partly by business rivalry, since Claude Code and Qoder have spent months competing for the same developers inside Alibaba as the most widely used productivity tools.

For the global market of AI-assisted coding tools, this is a sign that geopolitical tensions between the US and China are now directly shaping which code editor an ordinary developer is allowed to use. Companies like Alibaba are increasingly treating Western AI models as potential vectors for corporate espionage, regardless of whether a given mechanism actually had that intent.

Why it matters for Polish companies

For Polish companies using Claude Code or other Anthropic tools, the case brings no direct restrictions, the ban applies solely to Alibaba's internal policy. It's still worth paying attention to the mechanism itself, though, since it shows that popular AI coding tools can collect data about a user's environment, such as time zone or network settings, for purposes tied to protecting the vendor's intellectual property.

This is a good moment for security and legal teams at companies using AI coding assistants to review vendors' terms of service and privacy policies for what telemetry data is collected and why. The Anthropic-Alibaba dispute also shows that model providers are taking protection against distillation increasingly seriously, which could mean more similar monitoring mechanisms in popular tools going forward.

The case will likely keep developing, Anthropic has not withdrawn its distillation accusations, and Alibaba has neither confirmed nor publicly denied them. Both companies have a strong interest in presenting themselves as the victim, which makes it hard to verify the facts independently from outside.

Sources: Alibaba bans staff using Claude Code over Anthropic spyware concerns (scmp.com), Alibaba reportedly bans employees from using Claude Code (techcrunch.com), Anthropic Says Alibaba Used 25,000 Fake Accounts to Distill Claude (forbes.com).