An autonomous military artificial intelligence system developed in China outperformed human commanders in simulation tests. The system stood out with its speed and decision-making capabilities in command processes.
An autonomous military artificial intelligence system, dubbed a "digital chief of staff," developed by the Chinese People's Liberation Army (PLA) and the National University of Defense Technology (NUDT), was tested in field-like simulations. According to reports, the system was able to produce faster and more accurate decisions compared to human commanders in complex combat scenarios.
In simulations, the system, which focused particularly on high-intensity operations such as amphibious landings, demonstrated high performance even under conditions of severe communication disruption. The AI, reported to operate with over 90% recall accuracy, was able to maintain decision-making processes by preserving critical data despite disruptions in information flow.
Stood Out with Speed and Accuracy Advantage
The NUDT research team pitted the system against five senior military experts with an average of 12 years of experience. The test scenario was a high-intensity coastal landing operation, designed to resemble potential Taiwan conflict scenarios.
Throughout the simulation, the AI managed the complex command flow of troops advancing from the coast to inland areas, maintaining operational order. It was stated that the system took action much faster in moments when human commanders lost time evaluating options. According to the assessment, the AI accelerated the decision cycle, shortening the OODA (Observe, Orient, Decide, Act) process by 43%.
The system's core structure relies on combining large language models (LLMs) with real-time combat data. This structure enables the AI to filter operationally critical elements from intense information flow and present only decision-influencing data to commanders.
It was noted that the system specifically focuses on areas defined as "critical information requirements," meaning it identifies unknowns that determine the success of a mission and guides the command chain accordingly. This approach aims to prevent human commanders from overlooking details under stress.
Can Analyze the Fog of War
One of the most striking features of the AI during the simulation was its ability to detect missing information on the field. While analyzing enemy movements during the advance of armored units, the system identified the absence of some expected reserve units as a risk factor.
Following this detection, it was reported that the system recommended sending reconnaissance units without waiting for human intervention, thus allowing for an early assessment of a potential ambush scenario. It was stated that the system generated decisions in seconds, whereas human commanders required much longer discussions for similar assessments.
High Performance Despite Electronic Jamming
In one part of the tests, intense electronic jamming was applied, almost completely disrupting the digital combat environment. Despite this, the system's memory and data processing capabilities remained unaffected, and it continued to recall critical information with over 90% accuracy.
This feature demonstrated that the system could not only collect data but also maintain operational continuity even under disrupted communication conditions.
Although the results are remarkable, researchers emphasize that the system is not yet fully mature. One of the most significant limitations is defined as the "cold start" problem. If a commander is new or lacks a recorded decision history, the AI has no reference point to follow.
Furthermore, it is stated that at this stage, the AI focuses only on a specific area of expertise, having been trained particularly for coastal operations. This means it has not yet been extensively tested in different scenarios such as urban warfare or mountainous terrain combat.
It is anticipated that the developed system could evolve into a concept of "intelligent combat agents" managing multiple military units simultaneously in the future. This approach suggests that human intuition-based decision-making mechanisms on the battlefield may increasingly be replaced by data-driven systems.
0 Comments: