Chinese Military Researchers Use OpenAI and Anthropic Models to Advance Defense AI, Reuters Review Finds - Tekedia

Reuters found Chinese military-linked institutions using outputs from OpenAI and Anthropic models to train domestic defense AI systems.

The research relied on model distillation to build local tools for cyber operations, surveillance, and drone applications.

Examples in the review involved PLA-linked units, military universities, and defense academies.

The findings raise concerns that China may be speeding up military AI development while sidestepping U.S. export controls and intellectual property rules.