China Eliminates the Last Advantage Fighter Pilots Had Over AI: The Ability to Predict the Unpredictable

A recent study shows that AI technology can completely neutralize the human edge in real combat scenarios.

Fighter Jets
No comments Twitter Flipboard E-mail
miguel-jorge

Miguel Jorge

Writer
  • Adapted by:

  • Alba Mora

miguel-jorge

Miguel Jorge

Writer

Journalist. I've spent more than half of my life writing about technology, science, and culture. Before landing here, I worked at Telefónica, Prisa, Globus Comunicación, Hipertextual, and Gizmodo. I'm part of Webedia's cross-section team.

164 publications by Miguel Jorge
alba-mora

Alba Mora

Writer

An established tech journalist, I entered the world of consumer tech by chance in 2018. In my writing and translating career, I've also covered a diverse range of topics, including entertainment, travel, science, and the economy.

267 publications by Alba Mora

There was a small space within the military where “human power” continued to dominate over algorithms. In the realm of air power and the operation of fighter jets, pilots held an advantage that hindered the progress of AI: their unpredictable maneuvers. However, that advantage may soon diminish. Chinese engineers have discovered how to anticipate these movements.

The end of the human advantage. According to a recent study published in the Journal of Gun Launch & Control, China has made an unprecedented breakthrough in aerial combat. This advancement stems from the elimination of the last tactical edge that human pilots held over AI: the ability to execute unpredictable, high-intensity maneuvers to evade attacks.

Researchers at China’s Northwest Institute of Mechanical & Electrical Engineering, a key branch of arms supplier Norinco, have developed a system that combines advanced infrared imaging with AI-driven predictive modeling. This innovative approach accurately anticipates an opponent’s movements by detecting subtle adjustments to their control surfaces, such as ailerons and rudders.

The study. The Chinese breakthrough addresses a key limitation of AI systems in aerial combat. They depend on trajectory-based predictions, which struggle to respond effectively to the nonlinear and sudden moves of human pilots.

The research team focused on the physical mechanics of enemy aircraft. This enabled the AI to interpret millimeter-precise movements of control surfaces to predict actions before they occur.

How it works. The system is built upon a modified YOLOv8 neural network. This is the same technology that analyzes infrared images and detects micro-deformations in aircraft components, such as an F-15’s 5-foot rudder or its 6.5-foot stabilizer. This information is processed using a long- and short-term memory network enhanced with attention mechanisms. As a result, the technology can anticipate a pilot’s next move in real time without relying solely on previous flight patterns.

In simulations, the AI system successfully reduced prediction errors to less than 6.5 feet, which is 10 times better than traditional systems. When applied to automated anti-aircraft systems, this level of accuracy could enable direct attacks on the cockpit if necessary.

Fighter Jet

Testing. In their study, researchers evaluated the algorithm using flight profiles that simulated real combat tactics. In one scenario, an F-15 dropped munitions at low altitude and then executed a sharp climb, a maneuver that demands precise coordination of flight controls.

In another case, the aircraft performed rapid evasions, demonstrating erratic movements and abrupt direction changes. In both experiments, the AI system was able to predict the aircraft’s next moves milliseconds in advance, effectively overriding the pilot’s evasive maneuvers.

A new chapter in the arms race. This breakthrough arrives at a pivotal moment in the global competition for AI supremacy in aerial combat. In 2020, the Defense Advanced Research Projects Agency conducted the AlphaDogfight Trials, showcasing AI’s superiority in combat simulations. More recently, in 2024, an AI system participated in an actual dogfight against a human pilot for the first time.

What’s different now? The Chinese team highlights that its technology is designed to optimize fire control rather than enable fully autonomous attacks. However, the principles behind its AI model align with international programs such as the Air Force’s “loyal wingman” drones, which integrate piloted aircraft with combat AI systems.

Additionally, the study confirms that the new Chinese technology is feasible with current hardware. Chinese-made PS800 infrared sensors, which offer a resolution of 0.375 milliradians at distances of 5 miles, achieve the accuracy levels required for real combat operations.

Limitations and countermeasures. Despite its advanced capabilities, the system isn’t invulnerable. According to its creators, one of its main weaknesses is the potential interference from electronic and optical countermeasures.

For instance, an adversary could employ high-power lasers to blind or even destroy the infrared cameras that the AI relies on to detect enemy movements. This area still has huge room for improvement.

The future of aerial combat. The new Chinese AI technology marks a turning point in aerial warfare. Historically, human pilots held a critical advantage: their ability to perform unpredictable maneuvers to avoid being shot down.

However, the new algorithm can analyze even the most subtle movements of a fighter jet and predict its trajectory with remarkable accuracy. As such, human pilots may be on the verge of losing their final tactical advantage. As drones have already initiated changes in real combat scenarios, fully autonomous warfare is one step closer.

Images | Leo_Visions | Omar Barrera

Related | One of China’s Best-Kept Secrets, Revealed: This Is the New J-35A Fighter, a Rival to America's F-35

Home o Index
×

We use third-party cookies to generate audience statistics and display personalized advertising by analyzing your browsing habits. If you continue browsing, you will be accepting their use. More information