The Chinese Army trains an AI to reproduce the decisions of military leaders

It is common to be confronted with certain certainties minimizing the operational potential of the Chinese Army, when trying to assess Chinese military power, even from references on the subject.

If, in recent years, the discourse on the lack of performance and reliability of Chinese equipment has tended to fade, the lack of hardening of the PLA, a relatively young army, and less exposed to conflicts than the United States United or Europe since the end of the Korean War, has regularly advanced to attenuate the perception of effectiveness of Chinese forces.

The fact, even if the People's Republic of China is a much more bellicose nation than it tries to say (Tibet, Vietnam, Ladakh, etc.), the PLA is also aware of this weakness, and produces immense efforts to try to overcome it.

To achieve this, the Chinese armies organize numerous exercises, intra- and inter-arms, sometimes much more realistic than Western exercises with, in particular, live ammunition firing much more frequently than in the West. The other area of ​​improvement is based on the intensive use of simulations and wargames, to confront its officers with the decisions they may have to make.

To this end, the laboratory at the Joint Operations College of the National Defense University of Shijiazhuang, in Hebei province, has developed Artificial Intelligence designed to reproduce the behavior of military leaders, in the doctrine as in their biases, in order to further improve the realism of the wargames used to train PLA officers.

The Chinese Army wants to overcome the lack of realism in wargames

If Wargame has been intensively used for the training of Chinese officers and senior officers since the beginning of the 70s, the bodies in charge of these tools quickly became aware of the limits of the exercise.

Chinese army training
The People's Liberation Army organizes numerous realistic exercises to compensate for the lack of experience of its soldiers.

Indeed, the soldiers controlling the blue forces (for the PLA, the allies are designated red forces, and the adversaries, blue forces, unlike Western countries), had the same training as the trainees, and tended to use the units under their command, applying Chinese doctrines and strategies.

To try to overcome this problem, Chinese training centers then undertook to specialize some of their trainers, to reproduce the behavior of opposing officers. However, this approach, although more efficient in terms of simulation quality, also generates numerous constraints, in particular in the organization of the sessions themselves.

Indeed, depending on whether it was necessary to simulate engagements against the American, Korean, Japanese, Indian or even Vietnamese armies, it was necessary to have specialized trainers available in sufficient numbers, as well as a significant infrastructure, with, however, an efficiency all the more limited as the scale of the simulation is large, and therefore requires a high number of trainers.

The arrival of digital technologies, and more specifically Artificial Intelligence, opens many doors in this area.

People's Liberation Army AI replicates military leaders' decisions and cognitive biases

Let us remember, first of all, that the statutes of the PLA prohibit subordinating military action, whatever its level, to an automated decision, or to artificial intelligence.

Admiral Paparo Indopacom
Chinese AI must react like the PLA's potential adversaries. There is no doubt that one of them is qualified to reproduce the behavior of Admiral Paparo, who heads the Indo-Pacific command of the American armies.

There are 75% of this article left to read, Subscribe to access it!

Metadefense Logo 93x93 2 Artificial Intelligence | Defense Analysis | Military training and exercises

The Classic subscriptions provide access to
articles in their full version, and without advertising,
from 1,99 €.

For further



Last articles