US Army Develops AI-Powered Concealment and Detection Technologies

The US Army is conducting AI-based projects to detect enemy camouflage and defeat the technology. According to budget documents and official statements, these efforts could radically change how soldiers hide and find targets in the future.

US Army electronics engineer Grace Hummel stated that the algorithm succeeded in learning visual patterns of objects on its own, much like a soldier preparing for a vehicle recognition test. The developed system helps AI-powered military drones recognize camouflaged enemy forces by examining photo collections taken in different environments. The software scans live drone footage to flag potential targets for soldiers to review. The army recently tested this technology with a military unit and plans integration tests with drone swarms at Yuma Proving Ground in Arizona.

On the other hand, army budget documents reveal a separate effort aimed at countering these technological advancements. This initiative involves designing new camouflage patterns and developing experimental materials to confuse AI-powered drone hunters. The research aims to exploit the weaknesses of algorithms by developing countermeasures against target recognition patterns through artificial intelligence and machine learning. In practical application, this means producing unusual patterns that would cause a concealed soldier or tank to be perceived as another object.

This concept yielded promising results in a research project involving Carnegie Mellon University, Florida State University, and the Army Research Laboratory. In a study published in March, it was observed that after applying AI-generated camouflage to digitally altered images of vehicles, the computer's detection capability dropped from 99.5 percent to 38.2 percent. The report emphasized that effective camouflage must not only fool machine detectors but also remain convincing to human observers.