Chinese military researchers and tech giants used Claude to develop air-defense suppression tools targeting Taiwan

Chinese military researchers and tech giants have been found using Claude, a U.S.-developed frontier AI model, to code 16 air-defense suppression tools targeting Taiwan. The model was also used to draft specifications for anti-torpedo systems and fed 151 million training queries to Alibaba. This usage raises concerns about the dual-use nature of AI technologies and their potential impact on global security dynamics.

The discovery comes as China claims to have advanced AI models that may compete with those developed in the U.S. However, this case suggests that Chinese entities may be leveraging foreign AI tools to enhance their own defense capabilities. The use of Claude by Chinese actors underscores the growing reliance on international AI platforms for military and strategic purposes.

According to the evidence, 16 air-defense suppression tools were coded using Claude, which could be used to neutralize or disrupt enemy air-defense systems. Additionally, 151 million training queries were fed to Alibaba, indicating a large-scale effort to refine and optimize AI models for specific applications. These numbers highlight the scale and precision of the AI-driven activities being conducted by Chinese entities.

The implications of this usage are significant, as it may increase the cost of AI development for other nations, raise concerns about vendor lock-in, and complicate governance frameworks for AI technologies. The market may also react by tightening export controls or increasing scrutiny of AI tool usage by foreign entities.

This case is still developing, with ongoing investigations into the full extent of Claude's usage by Chinese actors. As AI continues to play a critical role in defense and technology, the incident underscores the need for robust oversight and international cooperation to manage the risks associated with cross-border AI deployment.