As militaries around the world actively explore the application of artificial intelligence (AI) in strategic planning, troop training, and frontline operations, the skills that military personnel need to possess have also changed. In order to enable future military generals and officers to better cope with intelligent warfare, Chinese military researchers recently proposed to redesign an AI course for graduate students with "teaching for war" as the core, and to deeply integrate actual combat scenarios into the teaching. This plan highlights that the Chinese military is actively cultivating military talents with practical capabilities in AI technology to cope with the changing patterns of future warfare.
Hong Kong's "South China Morning Post" reported on the 6th that this idea proposed by a research team at the School of Intelligent Science of the National University of Defense Technology of the Chinese People's Liberation Army was published in the July issue of the "Defence Industry Conversion" journal. Researchers pointed out that in the fall semester of 2022, the school will launch the "Artificial Intelligence Principles and Practice" course for the first time in science and engineering graduate students in non-AI related departments, aiming to help students with different backgrounds integrate emerging technologies with their own majors. However, they found that the case studies in the course lacked sufficient "military color" and must be completely reconstructed to better meet the practical needs of "teaching for war."
In terms of specific course planning, the research team proposed to be based on common unmanned combat scenarios and regard military operations as the main line connecting various AI theories. The course topics will accurately correspond to each stage of the military decision-making and strike process, covering from observing and assessing the battlefield situation, making decisions and taking actions, to finding, locating, tracking, engaging and finally evaluating the results of the battle.
For example, the "reinforcement learning" unit that explores how AI accumulates experience through trial and error will be taught through combat decision-making and war games; while the "uncertainty reasoning" that deals with incomplete or unreliable information can be practiced using fault diagnosis of unmanned systems and border intrusion warning systems. The research team hopes that this will enable students to internalize theoretical methods into "intelligent thinking" for solving military problems.
In order to strengthen practical skills, the researchers emphasized that courses should not only allow students to execute existing codes, but must provide clear tactical problems, data sets, algorithms and step-by-step tasks. One of the proposed exercises would require students to train multiple simulated drones to engage in coordinated operations to surround and lock on high-value targets. In addition, the team has also developed an online simulation platform that allows students to personally design maps and combat scenarios, configure troops and tasks, and train AI agents through simulated combat exercises. The platform also supports AI competitions between different student teams.
The National University of Defense Technology, directly affiliated with the Central Military Commission of the Communist Party of China, is China's top military science and technology education institution. In 2015, it was listed on the Entity List of Trade Restrictions by the U.S. Department of Commerce for allegedly using U.S. technology in supercomputers related to nuclear test explosions. While vigorously promoting the application of forward-looking technologies, the researchers who wrote the paper also clearly delineated the boundaries of military ethics, emphasizing that all exercises must consolidate an absolute principle, that is, AI is only used as a decision-making aid, and the final command and firing rights must still be completely in the hands of humans.
