Alibaba bans use of Anthropic’s Claude Code over alleged security risks

Alibaba has reportedly prohibited employees from using Anthropic’s popular AI coding assistant, Claude Code, inside the workplace, marking another escalation in the growing technological rivalry between leading American and Chinese artificial intelligence companies. The policy is expected to take effect on July 10, with employees instructed to transition to Alibaba’s internally developed coding assistant, Qoder, for software development tasks.

According to reports, the decision was driven by security concerns after developers alleged that Claude Code contained mechanisms capable of detecting details about users’ computing environments, including time zones, network settings, and proxy configurations. Some researchers claimed these features could potentially identify whether users were operating from China or connected to Chinese organizations.

Anthropic has rejected suggestions that the software contained a malicious "backdoor." Company representatives explained that the functionality was introduced as a temporary anti-abuse measure designed to detect unauthorized account resellers and reduce AI model distillation, a practice in which competing companies use the outputs of advanced AI systems to train their own models.

The company has stated that the feature was experimental and is expected to be removed in a future software update.

The latest restriction comes shortly after Anthropic publicly accused entities linked to Alibaba’s Qwen AI division of carrying out one of the largest known attempts to extract capabilities from Claude models through large-scale automated interactions. Anthropic alleged that millions of conversations generated through thousands of accounts were used to accelerate competing AI development, allegations that Alibaba has denied.

Industry analysts say the dispute reflects the increasingly competitive global AI landscape, where concerns over intellectual property, cybersecurity, and technological sovereignty are becoming as important as model performance itself.

Chinese technology companies have accelerated investment in domestic AI tools as geopolitical tensions and export restrictions continue reshaping international technology partnerships. At the same time, Western AI firms are strengthening safeguards designed to protect proprietary models from unauthorized use.

The latest developments highlight how the global AI race is evolving beyond innovation into a broader contest over security, regulation, and control of next-generation artificial intelligence technologies.