The US Air Force was the first to introduce AI, which controlled the U-2 reconnaissance aircraft. It was engaged in the search and identification of enemy missile launchers. The model was responsible for the combat systems onboard the aircraft, simulating navigator and scout.
The United States Air Force conducted a flight in which the AI acted as a co-pilot in a U-2 spy plane. The military department said that this is the first time that the model managed the military system independently. At first, the AI made more than a million training flights, but the Air Force notes that this is a leap for “technical development in the military field.”
A model called ARTUµ is the leading computer program that dominates chess, Go, and even video games without knowing the rules in advance. Despite the lack of scripts, the model could also fly the plane and make final decisions in difficult situations.
The AI carried out a reconnaissance operation during a simulated missile strike at the Beale Air Force Base. AI searched for enemy launchers and tracked enemy planes. As noted by the researchers, the success rate of the model reached 80%.
Military engineers noted that ARTUµ was useful during a specific simulation mission and helped pilots maintain an edge in decision making on the battlefield. The Air Force noted that this would help the fighters think about certain parts of the operation and concentrate on their direct responsibilities.
“Interacting with proven AI in all aspects of a conflict, and even temporarily assigning responsibility to it, can change the odds in our favor. Given the high stakes of global AI, superiority over science fiction should become our military norm, ”the military noted.