Chinese tech giant Alibaba has told employees to stop using Claude Code, the coding assistant made by US AI firm Anthropic. According to Business Today, Alibaba added Claude Code to its list of high-risk software following an internal review, with the ban taking effect across its offices from July 10.
The move comes amid a deepening feud between the two companies. According to Tom's Hardware, the ban followed the discovery of an alleged hidden "China-detection" backdoor in the software, and staff have been told to switch to Qoder, an Alibaba coding tool. The Indian Express reports that Anthropic is said to have covertly rolled out a version of Claude with embedded code designed to track whether users are based in China or affiliated with Chinese companies.
These claims about a backdoor remain allegations reported by the outlets rather than confirmed findings, and the exact technical details have not been independently established in the sources here.
There are signs the split may be mutual. France 24 frames the story as "Anthropic didn't want them using it anyway," suggesting the US firm had its own reasons to keep Chinese corporate users at arm's length. Anthropic has previously limited access to its products in China.
Multiple outlets, including The Indian Express, tie the episode to the broader escalation of the US-China AI rivalry, in which access to advanced American AI tools is increasingly restricted along geopolitical lines.
Why it matters: The dispute shows how quickly AI software is becoming a front in US-China tech tensions, with companies on both sides treating each other's tools as potential security risks and steering staff toward home-grown alternatives.