China warns of ‘backdoor’ risk in Anthropic’s Claude Code

A cybersecurity platform operated by China's Ministry of Industry and Information Technology issued a warning on Wednesday that it had identified a serious security "backdoor" risk in Anthropic's AI coding tool Claude Code, escalating technological tensions between the world's two largest economies as both sides scrutinize each other's AI tools.cnbc+2

The Chinese Warning

The alert, published through the National Vulnerability Database's WeChat account, said Claude Code contains a built-in monitoring mechanism capable of transmitting sensitive information — including users' geographic location and identity-related identifiers — to remote servers without user consent. The advisory applies to Claude Code versions 2.1.91 through 2.1.196, according to China's state-run Global Times.firstpost+2

The NVDB advised organizations and users to immediately review affected systems and either uninstall the impacted versions or upgrade to the latest release, in which the identified code has reportedly been removed. The agency also urged organizations to tighten controls on external network access for development tools and strengthen traffic monitoring across core business networks.storyboard18+1

The warning followed Alibaba's ↗ 10.45% decision last week to ban employees from using Claude Code for all work purposes, effective July 10. The company classified the tool as "high-risk software with security vulnerabilities" in an internal notice first reported by the South China Morning Post. Employees were instructed to switch to Qoder, Alibaba's in-house AI coding platform.tomshardware+2

Parallel U.S. Probe Into Chinese AI

The Chinese warning arrives as U.S. lawmakers pursue their own investigation into AI security risks flowing in the opposite direction. A House Committee investigation is probing the growing adoption of Chinese-built AI models by American companies, according to CNBC.techbuzz+1

The concern stems from the rapid uptake of models from firms like DeepSeek and Z.ai, which are seen as highly competitive with leading American rivals at a fraction of the cost. Open-source Chinese models can be "60% to 90% cheaper" than leading Anthropic and OpenAI models, Justin Summerville, who works on data and analytics at OpenRouter, told CNBC.cnbc

The joint investigation, announced in April by House Homeland Security Committee Chairman Andrew Garbarino and House Select Committee on China Chairman John Moolenaar, is examining the national security risks posed by foreign-built AI systems handling American corporate data.homeland.house

A Deepening Rift

The mutual accusations underscore a widening AI cold war. Anthropic already prohibits Chinese companies, as well as foreign entities owned by those companies, from using its models. Security researchers had alleged on Reddit and GitHub that Claude Code contained hidden code designed to detect whether users were based in China or affiliated with Chinese AI labs — claims that preceded both the Alibaba ban and the government warning.techcrunch+2

Meanwhile, U.S. AI firms OpenAI, Anthropic, and Alphabet's ↘ 0.82% Google have been working together through the Frontier Model Forum to detect adversarial distillation attempts by Chinese competitors, according to the Los Angeles Times.latimes