Chinese researchers tied to the People’s Liberation Army (PLA) have used Meta’s open-source Llama model to create a military AI tool called “ChatBIT,” aimed at intelligence gathering and decision support. A June paper by six researchers from PLA-affiliated institutions, including the Academy of Military Science, described adapting Meta’s Llama 2 13B model for military dialogue and information processing.
Meta, which provides Llama models publicly with restrictions, prohibits military and espionage uses but has limited enforcement options. In response to the findings, Meta’s policy director stated the PLA’s use violates company policies. The U.S. government is also monitoring these developments; recent AI-related policies reflect concerns about the national security risks of open AI models.
Researchers plan for ChatBIT to eventually support broader military applications, like strategic planning and training. Analysts say China’s rapid AI advancements make it challenging to limit their use of open-source technology, as collaborations between U.S. and Chinese scientists remain extensive.
