US developers to test combat artificial intelligence in real military scenarios

Saab has begun developing a unique system for assessing the effectiveness of artificial intelligence in combat. The project received support from the US Defense Advanced Research Projects Agency (DARPA) and aims to create a platform that will measure the effectiveness of AI in real-world military scenarios, Militarny reports.

According to Saab representatives, the main goal is to ensure transparency and accuracy of the assessment of decisions made by AI, as well as to test its ability to adapt to unpredictable combat situations. The system will feature the ability to simulate various scenarios, assessing how reasonable and effective the solutions proposed by the algorithms are.

The development is also focused on taking into account the ethical aspects of using AI in the military. Saab aims to create a system that will ensure not only combat effectiveness but also compliance with international norms and standards. This should reduce the risks of accidental losses and erroneous decisions.

The first prototypes of the system are expected to be ready for testing in 2026. They will be an important step in the introduction of modern technologies into the armed forces, contributing to increased efficiency and safety on the battlefield.

As a reminder, another American defense company, BlueHalo , demonstrated its advanced technologies at the Desert Guardian 1.0 event organized by the U.S. Central Command (CENTCOM) in Fort Drum, New York. Innovative solutions are aimed at combating unmanned aerial systems (UAS), in particular through the use of the VigilantHalo command and control platform.