Anthropic is pushing back against allegations from China that its AI coding tool, Claude Code, contains a security "backdoor."
According to Technology Org, Chinese authorities warned of a "backdoor" security risk in Claude Code, the AI-powered coding assistant made by the U.S. company Anthropic. The warning framed the tool as a potential security concern.
Anthropic has firmly denied the allegation. As reported by BankInfoSecurity, the company rejected China's claim that Claude Code contains a backdoor.
A "backdoor" is a hidden way to bypass a system's normal security controls, allowing outside access to data or systems without the user's knowledge. For a coding tool that developers weave into their software projects, such a flaw—if real—could carry serious implications. That is likely why the claim drew a quick and direct response from Anthropic.
The dispute lands against a broader backdrop of tension between the United States and China over artificial intelligence and the security of technology that crosses borders. Warnings about foreign software, and counter-denials from the companies behind it, have become a recurring feature of that rivalry.
The available sources report the core exchange—China's warning and Anthropic's rejection—but do not detail specific technical evidence for the backdoor claim or the full basis of Anthropic's denial.
Why it matters: when a government flags a widely used AI tool as a security threat and the maker flatly denies it, the standoff can shape which technologies companies and developers feel safe adopting across an increasingly divided global tech landscape.