Artificial Eyes: Generative AI in China’s Military Intelligence

Insikt Group assesses that the PLA and China’s defense industry are actively exploring and likely beginning to deploy generative AI for intelligence tasks, adapting foreign and domestic LLMs into specialized tools to process and analyze multi-INT data, generate products, and support decision-making; patents and procurement records indicate methods spanning OSINT, imagery processing, event extraction, and integrated military LLMs, with rapid uptake of domestic models like DeepSeek. While touted benefits include speed and scale, PLA sources acknowledge substantial risks—hallucinations, ideological bias, data reliability, and deepfake-enabled deception—leaving the depth and effectiveness of integration uncertain and posing technology transfer and counterintelligence challenges for the West.