China’s leading research institutions have developed an artificial intelligence tool for military purposes using the Llama model, which is publicly available by social media giant Meta. According to three academic papers and expert comments, these studies were conducted at research centers affiliated with the Chinese People’s Liberation Army (PLA).China uses Meta for military AI
According to an academic paper published in June, six researchers from the PLA’s Academy Military Science Center (AMS) developed an artificial intelligence tool called “ChatBIT” based on an early version of Meta’s Llama model. Meta’s Llama 13D large language model (LLM) was used in this study, and Chinese researchers added their own parameters, making the tool suitable for the purposes of collecting and processing military intelligence and providing reliable information for operational decisions.
The developed ChatBIT was optimized in the military field for dialogue and question-answer tasks. In the tests, it was claimed that ChatBIT showed superior performance than other artificial intelligence models, and even achieved a success close to OpenAI’s powerful ChatGPT-4 model. However, the researchers did not provide details about performance criteria or whether the model was actually used.
It was stated that this development is the first significant evidence that open-source large language models are being systematically researched for military purposes in China, especially by the PLA. While Meta continues to make its AI models available to the public, it prohibits the use of these models for purposes such as military or espionage. However, the fact that the models are publicly available means Meta has limited ability to enforce these rules.
China has invested at least $1 trillion in artificial intelligence
Meta clearly stated that this situation means a violation of the rules, and pointed out that China has invested more than 1 trillion dollars in the field of artificial intelligence. Meta said that while an AI built on legacy models is specifically unimportant, the real concern is that the US is not embracing open innovation.On the other hand, it is known that the researchers who developed ChatBIT are from AMS, Beijing Institute of Technology and Minzu University. It is stated that in the future, this artificial intelligence can be used not only in intelligence analysis, but also in areas such as strategic planning, simulation training and command decisions.
Experts argue that China is gradually closing the technology gap with the USA with its rapid progress in the field of artificial intelligence and its huge investments in this field. While Pentagon officials acknowledged that open source models have security risks as well as advantages, they stated that they will closely monitor China’s access to these technologies.
62425
