The Chinese tech firm Alibaba has put a ban in place for Anthropic’s Claude Code so that its employees can’t use it at work. The company is concerned that the tool can help to identify China-linked users, according to a source talking to Reuters. The ban comes after Anthropic accused Alibaba of extracting its Claude AI model’s capabilities without permission.
For those who don’t know, Claude Code is a tool made by Anthropic to help developers write code faster. Now that Alibaba has banned its use, it is telling its employees to use Alibaba Qoder, a rival service to Claude Code.
Toward the end of June, Anthropic hit out at Alibaba, claiming that the Chinese company had used 25,000 fraudulent accounts to extract the capabilities of Claude on a large scale. Anthropic claimed that the Qwen AI team has generated 28.8 million exchanges with Claude between April 22 and June 5 this year.
Anthropic says that this is not ordinary chatbot use. Instead, it says that it"s a concerted effort to collect answers that can be used to train or improve Alibaba’s AI models. Claude is known for being strong at coding, and this is one area that Alibaba allegedly focused on. It was also interested in learning from Claude’s multi-step reasoning and agentic abilities.
Alibaba was supposedly doing that for distillation purposes; this is where outputs from larger models are used to train smaller and cheaper models. What is noteworthy is the fact that Anthropic is now complaining about this unauthorized use, but it was also one of the many AI firms that pirated plenty of copyrighted content to develop its own AI models. It has even settled with various copyright holders, essentially admitting its fault.
Source: Reuters