The U.S. Air Force has quietly crossed a threshold in human‑machine teaming, with an F‑22 Raptor pilot taking direct control of an MQ‑20 Avenger drone from the fighter’s own cockpit during a mock combat mission. The test did more than prove a new data link, it showed that a front‑line stealth jet can manage a robotic wingman in real time while still fighting its own battle.
I see this as a pivotal moment in how air forces will fight, because it shifts drones from being remote assets on a separate network to extensions of the pilot’s own aircraft and tactics. Instead of a one‑to‑one relationship between a ground operator and a single unmanned system, the test points toward a future in which a single fighter crew could orchestrate a small formation of autonomous collaborators.
Inside the F‑22–MQ‑20 mock combat mission
The core of the demonstration was deceptively simple: a single F‑22 Raptor flew a mission while its pilot commanded an MQ‑20 Avenger that acted as a loyal wingman, sharing the airspace and the tactical picture. According to detailed accounts of the event, the fighter crew used an in‑cockpit interface to task the Avenger with surveillance and simulated targeting roles, treating the drone as another node in the formation rather than a separate asset controlled from the ground, a shift that earlier experiments had not fully achieved, as described in reporting on the F‑22 Raptor controls MQ‑20 Avenger test.
What stands out to me is how the mission blended routine flight operations with cutting‑edge autonomy instead of staging a one‑off science project. The MQ‑20 flew a realistic profile, maneuvering in coordination with the Raptor while the pilot issued commands that the drone’s onboard systems translated into specific actions, a level of integration that earlier surrogate tests had only hinted at, and which was further detailed in coverage of the mock combat mission.
How the pilot actually controlled the Avenger
The most consequential part of the experiment, in my view, was not that the MQ‑20 flew alongside the F‑22, but that the pilot managed it through a cockpit interface designed to minimize distraction from the primary job of flying and fighting the Raptor. Reporting on the trial explains that the crew used a dedicated control panel and display to assign tasks, monitor the drone’s status, and accept or refine autonomous recommendations, effectively turning the Avenger into a semi‑independent teammate rather than a remote‑controlled aircraft, a concept that was central to the combat flight test.
From what has been disclosed, the pilot did not fly the MQ‑20 with stick‑and‑rudder inputs, instead issuing higher‑level commands that the drone’s onboard autonomy executed within pre‑set safety and mission parameters. That approach aligns with the broader push toward collaborative combat aircraft, where human operators set intent and constraints while the unmanned systems handle the details, a design philosophy highlighted when General Atomics described the event as the first time an F‑22 pilot directly controlled a wingman drone from the cockpit in a public wingman drone briefing.
Why this test matters for collaborative combat aircraft
I see this F‑22–MQ‑20 pairing as a live‑fire rehearsal for the Air Force’s future collaborative combat aircraft, the family of drones meant to accompany crewed fighters into contested airspace. The Avenger served as a surrogate for those still‑in‑development platforms, giving engineers and operators a way to test how a pilot in a fifth‑generation jet can supervise an autonomous teammate without being overwhelmed, a role that was explicitly framed as a CCA stand‑in in coverage of the F‑22 pilot controls MQ‑20 demonstration.
What the test really validated, in my assessment, is the concept of mission‑level control, where the human sets objectives and priorities while the drone handles navigation, sensor management, and other lower‑order tasks. That is the only scalable way to imagine a single fighter crew coordinating multiple unmanned collaborators in a dense threat environment, and it is consistent with how insiders described the Raptor pilot collaborating with a drone wingman during the mock demo reported in an insider account of the event.
Mission roles the MQ‑20 can take off the pilot’s plate
One of the most practical outcomes of the trial is a clearer picture of what jobs a drone like the MQ‑20 can assume so the F‑22 pilot can focus on survivability and weapons employment. The Avenger’s payload capacity and sensor suite make it a natural candidate for forward sensing, target cueing, and even acting as a decoy or communications relay, roles that were explored when the pilot used the drone to extend the Raptor’s reach during the drone wingman flight.
In practice, that means the MQ‑20 can be pushed closer to suspected threats, soak up radar emissions, or loiter in higher‑risk areas while the crewed fighter remains in a more survivable position. I read this as a deliberate move toward a tiered risk model, where expendable or attritable unmanned systems absorb the brunt of the danger, a concept that General Atomics and Air Force officials have been refining through a series of MQ‑20 demonstrations, including the recent GA‑ASI MQ‑20 demonstration that emphasized modular payloads and flexible mission profiles.
What the test reveals about manned‑unmanned teaming doctrine
Beyond the technology, I see this event as a doctrinal marker for how the Air Force intends to fight with mixed formations of crewed and uncrewed aircraft. The mission showed that a single fighter can act as a local command node, delegating tasks to a robotic wingman while still integrating with broader command‑and‑control networks, a pattern that aligns with the service’s evolving manned‑unmanned teaming concepts described in coverage of the F‑22 and MQ‑20 MUM‑T test.
From a doctrinal standpoint, that suggests future packages could be built around a small number of high‑end crewed aircraft, each with several autonomous collaborators that can scout, jam, or strike in coordination with the lead. It also raises new questions about rules of engagement, data‑sharing authorities, and how much autonomy commanders are willing to grant to software agents in the heat of combat, issues that Air Force leaders have started to confront as they describe the MQ‑20 as a surrogate for collaborative combat aircraft in recent CCA surrogate drone discussions.
The road from one successful demo to operational reality
As impressive as this test was, I do not see it as a finished capability so much as a proof point on a longer path to operational use. Turning a one‑off demonstration into a fielded system will require hardening the data links against jamming, refining the cockpit interfaces so pilots can manage multiple drones under stress, and integrating the concept into training pipelines so new F‑22 and next‑generation fighter crews are comfortable commanding robotic teammates from day one, an evolution hinted at in several of the detailed accounts of the event and its follow‑on testing plans.
The bigger challenge, in my view, will be scaling this approach across platforms and theaters without creating brittle, over‑engineered systems that cannot adapt to real‑world friction. If the Air Force can keep the human at the center of the decision loop while letting autonomy handle the routine and the dangerous, the F‑22’s recent control of an MQ‑20 Avenger may be remembered as the moment when manned‑unmanned teaming stopped being a slide‑deck concept and started to look like an everyday part of combat aviation.
More from MorningOverview