On the morning of 1st December 2022, a modified F-16 fighter jet with the codename VISTA X-62A took off from Edwards Air Force Base, located approximately 60 miles north of Los Angeles. During a brief test flight, the VISTA engaged in advanced fighter maneuvers, including simulated aerial dogfights, all under the control of a sophisticated AI, marking the first time such AI was used on a tactical aircraft.
The US Department of Defense oversaw 12 AI-led test flights of VISTA X-62A between 1st and 16th December, totaling over 17 hours of autonomous flight time. These tests are part of the United States Air Force Vanguard’s initiative to develop unmanned combat aerial vehicles, known as the Skyborg program, which began in 2019 and is set to continue testing through 2023, with the aim of creating a working prototype by year-end.
M. Christopher Cotting, the director of research at the USAF Test Pilot School, emphasizes the significance of the VISTA program as a crucial first step toward these goals. He states that this approach, combined with focused testing on new vehicle systems as they are developed, will accelerate the advancement of autonomy for unmanned platforms, providing tactically relevant capabilities for military personnel.
The rise of autonomous combat systems, such as Ukraine’s use of semiautonomous drones and the US military’s first autonomous flight of a Black Hawk helicopter in the previous November, highlights the growing importance of autonomous technology in modern warfare. However, questions arise about the extent to which AI will dominate the skies and its implications for human pilots on the ground.
The VISTA X-62A, originally built in the 1980s and based on an F-16D Block 30 Peace Marble II, has a history of adaptability and versatility. It served as a simulation machine at the US Airforce Test Pilot School in the early 1990s, capable of mimicking the performance characteristics of various aircraft, from heavy bombers to light fighter jets. Before the recent autonomous flight tests, the VISTA received updates in the form of a “model following algorithm” (MFA) and a “system for autonomous control of the simulation” (SACS) from Lockheed Martin’s Skunk Works, as well as the VISTA Simulation System from Calspan Corporation, enhancing its autonomy and AI integration.
During the December testing, two AI programs, the Air Force Research Laboratory’s Autonomous Air Combat Operations (AACO) and the Defense Advanced Research Projects Agency’s (DARPA) Air Combat Evolution (ACE), were integrated into the system. AACO’s AI agents focused on beyond visual range (BVR) combat scenarios, while ACE concentrated on dogfight-style maneuvers with simulated visible enemies.
While a certified pilot remained in the rear cockpit of the VISTA as a backup during test flights, an engineer trained in AI systems occupied the front cockpit to address any technical issues that arose. Overall, these issues were minor and expected during the transition from virtual to live testing. This progress marked a significant step toward the Skyborg program’s goal of deploying autonomous aircraft in the near future.
The Department of Defense emphasizes that AACO and ACE are designed to complement human pilots, not replace them. AI copilot systems could serve as support mechanisms for pilots in active combat, capable of processing vast amounts of data and assuming control of the aircraft during critical moments. For routine missions that do not require human intervention, flights could be entirely autonomous, with the cockpit section of the plane replaced when a human pilot is unnecessary.
The objective is not to replace pilots but to augment their capabilities with AI tools. This augmentation can empower less experienced pilots with the advantages of AI assistance, leveling the playing field with more experienced counterparts.
Bill Gray, chief test pilot at the USAF Test Pilot School, views the integration of AI as a natural extension of pilot training. Gray believes that AI should be developed with safety measures to prevent costly mishaps, as an overreliance on AI can be perilous. He cites examples like Tesla’s autopilot program, which required drivers to remain attentive and at the wheel despite the autonomous capabilities. Cotting shares this perspective, emphasizing the importance of using the VISTA to test AI programs in a controlled environment to ensure safety.
The progress of AI integration in the USAF is rapid. Engineers have successfully switched autonomy algorithms onboard the VISTA in minutes during trial flights, demonstrating the adaptability of AI systems. However, the curriculum at the USAF Test Pilot School will require time to adjust to the new AACO and ACE platforms before students can fully utilize them in the VISTA cockpit.
While the broader military applications of AI remain undisclosed, image recognition technology incorporating AI is already prevalent across the military. While AI-driven tanks may not be on the immediate horizon, the skies are poised to witness the rise of a new form of intelligence.
Leave a comment