Using a touchscreen pilot interface, a human “battle manager” in the cockpit of a flying aircraft gave real-time commands to AI-controlled aircraft.

Skunk Works, Lockheed Martin’s secretive advanced projects division, along with its demonstration and prototyping team and the University of Iowa’s Operator Performance Lab (OPL), recently completed an uncrewed team mission. A human “battle manager” in the cockpit of an airborne aircraft used a touchscreen interface to issue real-time commands to AI-controlled planes.

During flight tests, the teams simulated an air combat mission: the human fighter commander aboard an L-39 Albatros assigned targets to two AI-powered L-29 Dolphin jets. The AI-controlled jets worked together to take on two fake enemy planes, using virtual mission systems and weapons to accomplish the task.

John Clark, director of Lockheed Martin Skunk Works, said:

“This work with the University of Iowa’s OPL lays the groundwork for the future of air combat, where manned and unmanned systems will work together for complex missions. It’s exciting to bring together so much talent to push the boundaries of this new approach.”

These tests build on previous attempts that focused on AI handling tasks such as air-to-ground jamming and target localization. This time the focus shifted to air-to-air combat, with the AI ​​directly controlling the plane via their autopilot system. It is the third test of its kind but the first with a human monitoring the AI ​​in real time.

Skunk Works focuses on developing systems where humans and AI-powered aircraft can join forces, improving mission flexibility, speeding decision-making and improving pilot safety.

Remarkably, the two L-29 Delfin jets involved in the tests were probably the same two nearly 60-year-old aircraft, deployed to Edwards Air Force Base earlier this year to support the operations of the United States Air Force Test Pilot School and collect data for ongoing research in machine learning and autonomy.

As we reported in detail last month, the OPL L-29 Delfin jet, originally designed in the late 1950s (the first flight of the type dates back to April 5, 1959), has undergone significant modifications to meet modern test requirements. Equipped with advanced displays, experimental data links and an array of sensors, these aircraft act as flying laboratories that can rapidly collect and analyze data.

Crewed and unmanned teaming

The concept of unmanned teaming and the role of AI in advancing air combat tactics is a recurring topic we’ve talked about often here at The aviator. Many articles highlighted the significant progress made in integrating autonomous systems with human-piloted aircraft to improve mission success and operational flexibility.

The “Loyal Wingman” concept typically involves unmanned or autonomous aircraft acting as supporting, AI-powered wingmen to human-piloted aircraft, performing missions and tasks in close coordination with their human counterparts. In more recent parlance, the preferred concept is CCA (Cooperative Combat Aircraft), which represents a highly integrated, collaborative approach to air combat, where multiple manned and unmanned aircraft work together as part of a cohesive system, often with AI-powered assets to perform complex missions along with human operators. In more general terms, the term MUM-T (Manned-Unmanned Teaming) is still used to refer to different types of manned and unmanned systems working together, with flexibility in mission and operation.

The scenario of the latest test, where a “battle manager” in the cockpit used a touchscreen interface to control AI-controlled aircraft, is consistent with the Loyal Wingman concept: the idea is to integrate AI-driven systems (the Loyal Wingman aircraft) with human pilot platforms, where the AI ​​systems can perform assigned tasks, such as surveillance, combat or tactical operations, under the direction of the human operator. This improves mission flexibility, reduces risks to human pilots, and improves overall mission efficiency by leveraging AI for faster decision-making.

Recently, we reported on how Lockheed Martin’s Skunk Works is a driving force in integration of CCA with crew platforms. These CCAs act as “autonomous wingmen” and provide capabilities such as reconnaissance, electronic warfare and even direct engagement with enemy targets.

“By using CCAs as a force multiplier, manned aircraft can effectively serve as a much larger force than they already are. The unmanned platforms can be used as simple ‘missile trucks’, allowing a larger payload to be transported into the battlespace, or more advanced programming could enable the drones to operate further away from their control aircraft and extend the force’s situational awareness and engagement range.Electronic warfare packages could also be carried to allow the CCA to act as decoy without adding additional risk to human pilots.

Considered a replacement for the F-22, the NGAD is usually described as a “family of systems,” including the manned aircraft; autonomous, unmanned “wingmen” drones for ISR, decoy or strike missions, and other disaggregated capabilities. These drones, called CCAs (Collaborative Combat Aircraft) in American language, are considered part of the family of systems.”

This change is seen as essential for adapting to modern warfighting requirements, where threats evolve rapidly, requiring advanced coordination between AI-powered platforms and human operators. The article notes:

Called a ‘family of systems’, NGAD emphasizes operational agility and adaptability. Autonomous drones and AI systems are central to this strategy, where the collaboration between human pilots and AI is designed to outpace the decision-making speed of adversaries.”

Another article dive deeper into the application of AI in tactical scenarios, especially in air-to-air combat. It describes how AI enables autonomous aircraft to handle dynamic mission objectives, such as engaging multiple threats while responding to real-time changes on the battlefield.

“AI-powered systems reduce pilot risk while allowing faster decisions on the battlefield. The technology enables drones to perform maneuvers and engage targets without waiting for human commands, while leaving ultimate control in the hands of the human pilot.”

The adaptability of these systems is further emphasizedespecially their ability to work seamlessly in contested environments.

Either way, it is interesting to observe how these innovations are now being tested and refined to ensure operational readiness in the future, with a strong focus on minimizing human risks while optimizing mission outcomes.

H/T to our friend Ryan Chan for heads-up!