In a move that underscores the deepening technological rift between Beijing and Silicon Valley, China’s Ministry of Industry and Information Technology (MIIT) has issued a high-level security warning regarding Claude Code, a sophisticated AI programming tool developed by the American firm Anthropic. The National Vulnerability Database (NVDB) claims to have detected hidden monitoring mechanisms within versions 2.1.91 through 2.1.196 of the software, labeling the threat as a serious risk to national data security.
According to the MIIT alert, the tool is accused of exfiltrating sensitive user information, including geographic data and identity markers, to remote servers without explicit consent. The regulator has instructed domestic institutions and developers to immediately purge the affected versions or upgrade to cleared alternatives, while calling for more aggressive traffic monitoring on core business networks to prevent unauthorized data outflows to foreign entities.
This security designation arrives at a critical juncture for the global AI industry, as autonomous agents begin to play a central role in software engineering. While Anthropic has built its global reputation on 'AI safety' and ethical alignment, Beijing’s characterization of its coding tool as a potential espionage vector suggests a new phase of the 'AI Cold War' where even developer-focused utilities are scrutinized as national security threats.
The regulatory pressure is already reshaping the domestic landscape, with Chinese developers being nudged toward homegrown alternatives. In the wake of the MIIT’s warning, domestic players such as DeepSeek have moved quickly to promote their own solutions, highlighting a coordinated effort to insulate China’s software development lifecycle from Western-controlled AI models and infrastructure.