Chinese Military Researchers and Tech Giants Used US AI Model Claude for Defense Applications, Report Alleges
A recent investigation has revealed that Chinese military researchers and technology companies allegedly used Anthropic's Claude AI model—a US frontier-level system—to support defense-related work.
According to the findings, the users reportedly developed at least 16 air-defense suppression tools specifically targeting Taiwan, along with anti-torpedo specifications. Additionally, the reports indicate that approximately 151 million training queries were processed using Alibaba's AI infrastructure.
The use of US-developed frontier AI models by foreign military entities raises questions about access controls, export considerations, and the boundaries of AI deployment in sensitive defense applications. Anthropic, the developer of Claude, maintains terms of service designed to prevent misuse for weapons development and other harmful purposes.
This situation highlights ongoing tensions between AI accessibility and national security concerns, particularly as advanced language models become increasingly capable and their deployment across international borders continues to grow.