Alibaba bans Anthropic's Claude Code after an alleged hidden China-detection backdoor is uncovered — employees told to switch to Qoder as the rift between the firms widens - Tom's Hardware

Alibaba has taken the unusual step of banning Anthropic's Claude Code among its employees after an internal security review uncovered an alleged hidden backdoor mechanism. The tool, designed to detect whether users are located in China, was flagged as a potential privacy and compliance risk, prompting Alibaba to instruct its workforce to switch to Qoder, its own AI coding assistant. The move underscores growing tensions between the two companies, as Alibaba accelerates its push to reduce reliance on foreign AI tools amid geopolitical and regulatory pressures.

The alleged backdoor was reportedly embedded in a way that triggered automatic detection of Chinese IP addresses or device settings, raising concerns about data sovereignty and unauthorized surveillance. Anthropic has not yet publicly addressed the claims, but the incident adds to a broader pattern of scrutiny over AI tools' compliance with regional data laws. Alibaba's decision reflects its broader strategy to prioritize homegrown solutions like Qoder, which aligns with its compliance frameworks and business interests in China.

Source: Alibaba bans Anthropic's Claude Code after an alleged hidden China-detection backdoor is uncovered — employees told to switch to Qoder as the rift between the firms widens - Tom's Hardware. Read the full piece at the source.