Chinese researchers with ties to China’s People’s Liberation Army (PLA) have developed an AI model called ChatBIT, designed specifically for military applications. The model has been built by researchers associated with the Academy of Military Science (AMS), the PLA’s top research group. According to several academic papers and analysts, ChatBIT uses a large language model (LLM), which has been modified for intelligence gathering and operational decision-making.
These papers highlight that ChatBIT is optimized for dialogue and question-answering tasks within the military field. It is reported that ChatBIT performs at approximately 90% of the level of OpenAI’s GPT-4 LLM, although there are no details on the specific testing methods or evidence of field use. The open-source nature of the model used for ChatBIT makes enforcement difficult.
ChatBIT’s development for PLA applications
It has been noted that the version used is an outdated one, emphasizing the rapid advancements China is making in AI technologies. Some experts have raised questions regarding the viability of ChatBIT, as it was trained using only 100,000 military dialogue records, a data set much smaller than what is typically used for training advanced models.
This suggests that ChatBIT might be a proof of concept, with plans to develop more extensive models in the future. The development of ChatBIT and its military implications have caused concerns in Washington, where policymakers are wary of open-source technologies being leveraged by adversaries to gain a technological edge. Sunny Cheung, an associate fellow at the Jamestown Foundation, a Washington, D.C.-based think tank, commented, “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 for military purposes.”
While the use of open-source technology can drive technological advancements, it also presents a dilemma for American policymakers who do not want to inadvertently aid their opponents.
This situation underscores the complexities and challenges of maintaining technological leadership in an era of rapid and open technological proliferation.
Cameron is a highly regarded contributor in the rapidly evolving fields of artificial intelligence (AI) and machine learning. His articles delve into the theoretical underpinnings of AI, the practical applications of machine learning across industries, ethical considerations of autonomous systems, and the societal impacts of these disruptive technologies.
