Alibaba bans employees from using Claude after security concerns and alleged backdoor risks sparked a new clash with Anthropic.

Alibaba bans employees from using Claude amid security concerns

Alibaba bans employees from using Claude in a move that shows just how tense the global AI race has become. The Chinese tech giant has reportedly prohibited staff from using Anthropic’s Claude Code at work, following concerns that the tool contained hidden mechanisms that could identify China-linked users and potentially expose sensitive information.

The ban is set to take effect on July 10, and Alibaba has reportedly told employees to remove Claude models from work devices and switch to its own coding platform, Qoder. That makes the decision more than a simple software policy change — it is now part of a wider strategic split between two companies that are already in direct conflict.

Why Alibaba made the move

According to reporting from Reuters and the South China Morning Post, Alibaba’s internal notice classified Claude Code as high-risk software after a security review. The company’s concern is tied to claims that Anthropic embedded code in Claude Code that could detect whether a user was in China or connected to Chinese AI labs.

The reported detection checks allegedly looked at things like proxy settings and time zones, and then matched them against a hidden list of Chinese networks and AI companies. Anthropic has not publicly issued a detailed response, but a Claude Code team member reportedly said the mechanism was intended to prevent account abuse and model distillation, and would be removed in a future release.

Why this matters

This is not just a workplace software ban. It is part of a much deeper dispute over trust, competition, and control in AI. Anthropic has already accused Alibaba of illegally extracting Claude capabilities, which means both companies are now treating the other’s technology as a security concern rather than a neutral tool.

That kind of escalation matters because AI coding assistants are increasingly embedded into real business workflows. When a company believes a tool may be watching user environments too closely, the reaction is usually swift, especially if sensitive code, infrastructure, or data might be involved. In this case, Alibaba appears to have chosen caution over convenience.

What employees are being told to use

Alibaba is reportedly steering staff toward Qoder, its own coding platform, as the alternative. That move gives the company more control over security policies, data handling, and internal compliance. It also keeps employees inside Alibaba’s own AI ecosystem rather than depending on a U.S.-based competitor.

That is a familiar pattern in enterprise technology. Once a company decides a third-party AI tool is too risky, it usually replaces it with an in-house or more tightly governed option. The difference here is that the ban lands in the middle of a highly visible geopolitical and commercial fight over AI leadership.

Bigger AI industry signal

Alibaba’s decision is a reminder that AI adoption is no longer just about model performance. Security, provenance, and visibility into what the software is doing behind the scenes are becoming just as important. As AI tools become more powerful, companies are also becoming more suspicious of how those tools gather data, classify users, and interact with sensitive systems.

This episode may also push other enterprises to scrutinize their own AI coding assistants more carefully. If a tool can inspect environment details, identify regional usage patterns, or insert hidden markers into prompts, even a temporary controversy can quickly become a policy problem.

Final take

Alibaba bans employees from using Claude because it sees the risk as more than theoretical. Whether the issue is about hidden code, geopolitical friction, or a broader battle for AI dominance, the message is clear: in today’s AI race, trust can disappear fast.

Summary: Alibaba has reportedly banned Claude Code at work over security concerns, allegedly linked to hidden detection mechanisms and a growing dispute with Anthropic over AI trust and model extraction.