Chinese Researchers Use Meta’s Llama Model For Military Applications

Multiple Chinese research institutions have been using Meta’s public advanced AI model Llama to develop military AI tools. According to an exclusive shared by Reuters today, these institutions are linked to the People’s Liberation Army (PLA).

Analysts from the Academy of Military Science (AMS) and the PLA shared with the news agency details and evidence. One of the papers, reviewed by Reuters in June, showed how the academic institutions had been using Meta’s early large language model (LLM) Llama 13B to develop a military tool called “ChatBIT.”

“It’s the first time there has been substantial evidence that PLA military experts in China have been systematically researching and trying to leverage the power of open-source LLMs, especially those of Meta, for military purposes,” said Sunny Cheung, an associate fellow specializing in China’s dual and emerging technologies at the Jamestown Foundation.

ChatBIT gathers and processes information to help in strategic decision-making. According to the paper, it had been “optimised for dialogue and question-answering tasks in the military field,” and could outperform other AI models like ChatGPT. Researchers said ChatBIT could be applied for strategic planning, simulations, and other situations in the future.

In another paper, two researchers from the Aviation Industry Corporation of China (AVIC)—a firm linked to the PLA by the U.S. government—reported using model LLama 2 for “the training of airborne electronic warfare interference strategies”.

Meta’s LLM isn’t entirely open-source as it has restrictions to prevent misuse. While its terms explicitly prohibit the use of the model for military purposes, the tech giant has limited control over its public model.

“Any use of our models by the People’s Liberation Army is unauthorized and contrary to our acceptable use policy,” said Molly Montgomery, Meta’s director of public policy to Reuters.

Another spokesperson from Meta considered the use of old Llama models “irrelevant” as they are certain that China is already investing over a trillion dollars to win the AI race and surpass the U.S. in technological developments.

The technology competition between China and the U.S. has intensified their friction. A few days ago, the U.S. government announced the final details of the new rules to limit American investments in certain Chinese technology industries—specifically in AI.