Top Chinese research institutions connected to the People’s Liberation Army (PLA) have leveraged Meta’s publicly available Llama model to create an artificial intelligence (AI) tool with potential military applications, according to a review of academic papers and analysis from experts.
In a June paper examined by Reuters, a team of six Chinese researchers from three institutions—including two affiliated with the PLA’s Academy of Military Science (AMS)—described how they utilized an early version of Meta’s Llama to develop what they have named “ChatBIT.”
The researchers employed an earlier iteration of the Llama 2 13B large language model (LLM) from Meta, incorporating their parameters to construct a military-oriented AI tool designed for intelligence gathering and processing, aimed at delivering accurate and reliable information for operational decision-making.
According to the paper, ChatBIT was fine-tuned and optimized for dialogue and question-answering tasks relevant to military operations, reportedly outperforming some other AI models that were approximately 90% as capable as OpenAI’s ChatGPT-4. However, the researchers did not clarify how they defined performance or whether the AI model has been deployed operationally.
“This marks the first substantial evidence of PLA military experts in China systematically researching and seeking to leverage open-source LLMs, particularly those from Meta, for military purposes,” stated Sunny Cheung, an associate fellow at the Jamestown Foundation who focuses on China’s emerging and dual-use technologies, including AI.
Meta has actively embraced the open release of many of its AI models, including Llama, but imposes restrictions on their use. These include requiring services with over 700 million users to obtain a license, and prohibiting applications related to military operations, warfare, nuclear industries, espionage, and other activities subject to U.S. defense export controls.
Despite these restrictions, the public nature of Meta’s models limits the company’s ability to enforce compliance.
In response to inquiries from Reuters, Meta reiterated its commitment to its acceptable use policy and stated that it has implemented measures to prevent misuse. “Any use of our models by the People’s Liberation Army is unauthorized and contrary to our acceptable use policy,” Molly Montgomery, Meta’s director of public policy, told Reuters in a phone interview.
Meta further emphasized the importance of open innovation in the context of global AI competition. “In the global competition on AI, the alleged role of a single, outdated version of an American open-source model is irrelevant when we know China is already investing over a trillion dollars to surpass the U.S. in AI,” a Meta spokesperson noted in a statement.
The team of researchers includes Geng Guotong and Li Weiwei from AMS’s Military Science Information Research Center, as well as researchers from the National Innovation Institute of Defense Technology, Beijing Institute of Technology, and Minzu University. They indicated that, with further technological refinement, ChatBIT could eventually be applied not only to intelligence analysis but also to strategic planning, simulation training, and command decision-making.
China’s Defense Ministry did not respond to a request for comment, nor did any of the involved institutions or researchers.
While Reuters could not independently verify ChatBIT’s capabilities and computing power, the researchers acknowledged that the model was built using only 100,000 military dialogue records—a relatively small dataset compared to other LLMs trained on trillions of tokens. Joelle Pineau, a vice president of AI Research at Meta and a computer science professor at McGill University, expressed skepticism regarding the model’s effectiveness. “That’s a drop in the ocean compared to most of these models… it really makes me question what they actually achieve here in terms of different capabilities,” she said.
This research emerges amid ongoing debates in U.S. national security and technology circles regarding whether companies like Meta should make their models publicly available. In October 2023, U.S. President Joe Biden signed an executive order aimed at managing AI developments, emphasizing both the substantial benefits and significant security risks associated with open-source models.
This week, Washington announced it was finalizing rules to limit U.S. investment in AI and other technological sectors in China that could pose national security threats. Pentagon spokesman John Supple acknowledged that while open-source models have their advantages and disadvantages, “we will continue to closely monitor and assess competitors’ capabilities.”
Some analysts argue that China’s advancements in indigenous AI, including the establishment of numerous research labs, have made it increasingly challenging to keep the country from closing the technology gap with the United States. In a separate academic paper reviewed by Reuters, researchers from the Aviation Industry Corporation of China (AVIC)—an entity with ties to the PLA, according to U.S. designations—described using Llama 2 for developing airborne electronic warfare interference strategies.
China’s adoption of Western-developed AI technologies has also permeated domestic security, as indicated in a June paper that discussed utilizing Llama for “intelligence policing” to analyze large datasets and enhance police decision-making processes.
The state-run PLA Daily published commentary in April discussing how AI could expedite the research and development of military equipment, improve combat simulations, and enhance training efficiency.
“Can you keep them (China) out of the cookie jar? No, I don’t see how you can,” William Hannas, a lead analyst at Georgetown University’s Center for Security and Emerging Technology (CSET), remarked to Reuters. A 2023 CSET paper identified 370 Chinese institutions whose researchers published works related to General Artificial Intelligence, contributing to China’s strategy to dominate the global AI landscape by 2030. “There is too much collaboration between China’s leading scientists and the U.S. AI researchers for them to be excluded from developments,” Hannas concluded.
MUST READ:
https://flarenews.pk/2024/11/02/constantine-departs-after-leading-pakistan-to-historic-success/
