AI fighter jet fleet attack from the USA
Earlier this month, the U.S. Department of the Air Force reported flying the 9.15-meter-long XQ-58A Valkyrie drone for three hours in a first-of-its-kind fully AI-controlled flight test. Now, the US Air Force is requesting approximately $6 billion to build a fleet of up to 2,000 Valkyrie aircraft; Each aircraft costs an estimated $3 million, much cheaper than human-operated F-35 or F-22 fighter jets.
Congress must approve the five-year project to build the drone fleet, which is projected to cost $5.8 billion over that period, with $3.03 billion in fiscal year 2028 alone. However, the US Air Force describes this project as an effort to develop a “Cooperative Combat Aircraft … capable of developing crewed weapon systems to achieve air superiority.”
“Air Force unmanned aerial vehicles will be designed to allow commanders and operators to exercise appropriate levels of human judgment regarding the use of force,” a Pentagon spokesman said. Therefore, even if these vehicles work with artificial intelligence, they will need human approval to fire or engage the weapon.
The planned mission scope will reportedly be tested later this year in a mission assessing the Valkyrie’s ability to track and destroy an enemy target in a simulated scenario over the Gulf of Mexico. The experiment will also put the aircraft’s AI flight software into a dogfight scenario. The Air Force wants to contract with private businesses like Kratos or Shield AI to develop software and hardware components separately. According to officials’ estimates, it could take five to ten years to produce an AI drone capable of real-life aerial combat.