Chinese military researchers caught using Claude for air-defense and weapons (www.tomshardware.com)

Recent disclosures from Anthropic reveal that Chinese military researchers have been using the AI model Claude for various military and surveillance initiatives, marking a significant development in the intersection of AI technologies and defense strategies. Notably, two military projects involved drafting specifications for an anti-torpedo system, and creating software modules for electronic warfare that targeted sensitive locations in Taiwan. The use of Claude, rather than domestic alternatives, suggests a preference for its superior capabilities in coding and reasoning, potentially leveraging the model's vast information on U.S. military technology. In addition, Anthropic reported that Chinese-linked entities employed Claude for extensive surveillance efforts against Uyghur communities, facilitating infiltration and mapping of social networks among diaspora groups. These findings highlight not only the capabilities of Claude in critical applications but also the concerning trend of AI models being co-opted for strategic military and surveillance purposes. Furthermore, the report revealed that prominent Chinese companies, including Alibaba and Xiaomi, have engaged in large-scale "distillation" efforts to extract Claude's capabilities, indicating a robust and ongoing technology transfer landscape that poses challenges for AI governance.