Chinese researchers have developed a military commander artificial intelligence ‘caged’ that is based on the best strategic planners living and dead and designed to reflect a human commander in all aspects, from experience to thought patterns and personality and even his mistakes.
According to the South China Morning Post, the commander AI is confined to a laboratory at the Joint Operations College of the National Defense University in Shijiazhuang, Hebei Province. According to its creators, who revealed the project in a peer-reviewed article published in May in Common Control & Simulation, the AI has already participated in large-scale war games for the People’s Liberation Army, PLA. These involve all branches of the EPL and in them the AI has had unprecedented authoritylearning and evolving while carrying out constant virtual wars.
The team, led by principal engineer Jia Chenxing, notes that artificial intelligence has both potential and risks in military applications in real life and that this project offers a viable alternative for its improvement.
In China, the army must strictly abide by the principle that ‘the Party rules the gun’; that is, only the Central Military Commission of the Communist Party of China can mobilize the PLA. But now AI has the power to exercise that same authority within the confines of the laboratory, without human interference. This is the highest-level publicly reported role of an AI in military research.
Researchers point out that the EPL has developed numerous operational plans for possible military conflicts in regions such as Taiwan and the South China Sea. Jia and his colleagues say scientists should test these plans in simulations to ‘weigh the good and the bad and gain insight into the chaos of battle.’
Military simulations often require the involvement of human commanders to make immediate decisions in response to unexpected events. But the number of senior PLA commanders and their availability is very limitedwhich makes it impossible for them to participate in a large number of war simulations, the media notes.
Jia’s team explained that the initial configuration of the commander AI resembles that of an experienced and brilliant strategist. They assure that AI has ‘solid mental faculties, a calm and firm character, and the ability to analyze and judge situations calmly.’
They also mention that AI lacks emotional or impulsive decisions and is quick to come up with practical plans by recalling similar decision-making scenarios.
‘The personality of the virtual commander can be adjusted if deemed necessary‘, they add. Under immense pressure, individuals can struggle to make completely rational decisions within strict time constraints, according to Jia’s team. Instead of using pure analysis, commander AI increasingly relies on empirical knowledge for its combat decisionssearching for satisfactory solutions, remembering similar scenarios and quickly formulating a viable plan.
AI commander allows PLA conduct a large number of war simulations ‘without humans in the loop’. Identify new threats, make plans and make decisions based on the overall situation when battles fail or results are not as expected. He also learns and adapts from wins and losses.