Leading Chinese research institutions affiliated with the People’s Liberation Army (PLA) have used Meta’s Llama large-scale language model to create artificial intelligence-based military tools. Chinese researchers have developed a ChatBIT tool designed to collect, process intelligence and provide accurate information for operational decisions, Mezha.Media reports.
The articles, which were accessed by journalists, indicate that six Chinese researchers used an early open-source version of the Llama 13B model, adding their own parameters to optimize the model for military purposes. Two of the authors of the study were from the Academy of Military Sciences, which is part of the PLA. ChatBIT, according to the researchers, can outperform 90% of AI models in military tasks, approaching the power of OpenAI’s GPT-4.
Meta prohibits the use of its models for military activities, weapons development, the nuclear industry, and other industries subject to US export controls. A Meta spokesperson said that any use of the PLA models is unauthorized and contrary to company policy. In a telephone comment to Reuters, Meta representatives emphasized that China is actively investing in AI development, and the cost of its programs reaches trillions of dollars, which is a serious competition for the United States.
According to the documents, ChatBIT has prospects for development as a tool for strategic planning, simulation training, and team decision-making. No official comments have been received from the Chinese Ministry of Defense or from institutions and researchers.
Leave a Reply