China warns about AI risks with Anthropic's Claude Code

In a significant development, China has issued a warning regarding potential security issues linked to Anthropic's Claude Code, an artificial intelligence tool developed by the U.S.-based company. This alert comes amid escalating tensions in the technology sector between the U.S. and China. The concerns stem from accusations made by Anthropic last month, which claimed that Chinese tech giant Alibaba was attempting to gain unauthorized access to its AI capabilities that remain unavailable in China. While Alibaba has refrained from responding to these allegations, the interest in U.S. AI tools among Chinese users continues to grow. A Xiaomi AI developer noted at a state-sponsored forum in March that many individuals in China are actively utilizing Claude Code. In a recent statement, China's Ministry of Industry and Information Technology highlighted findings from its cybersecurity threat platform, asserting that the Claude Code tool contains a 'back-door' vulnerability. This flaw poses a significant risk, as it could potentially allow sensitive data, including user identities and locations, to be transmitted to external servers without user consent. To mitigate these risks, users are advised to either uninstall or upgrade their versions of Claude Code, specifically those between versions 2.1.91 and 2.1.196, which were released from April 2 to June 29. According to Anthropic's website, the most recent version available as of Wednesday is 2.1.204. As of now, Anthropic has not provided a response to these serious security concerns.