It’s Easier to Replace Fighter Pilots with AI Than Aircraft Mechanics

Fighter pilots might want to start updating their résumés. Aircraft mechanics can probably relax for a while.

That may sound like a ridiculous prediction, especially considering the skill, training, judgment, and experience required to fly a modern combat aircraft. But recent developments in artificial intelligence suggest we may be approaching an extraordinary technological milestone: replacing the person flying a fighter jet could prove considerably easier than replacing the people who keep it flying.

And as a former Air Force aircraft maintainer, I must admit there’s something deeply satisfying about that possibility.

For decades, fighter pilots have occupied the glamorous end of military aviation. They get the movies, the recruiting posters, the aviator sunglasses, and the admiration of people who apparently believe airplanes maintain themselves.

Aircraft mechanics get hydraulic fluid in their hair, safety wire embedded in their fingers, and the opportunity to explain why a jet that was perfectly serviceable yesterday has suddenly developed three unrelated discrepancies.

Now artificial intelligence threatens to turn that traditional hierarchy upside down.

The Air Force may eventually replace the guy in the cockpit with a computer while discovering that the oil-stained mechanic standing underneath the airplane is considerably more difficult to duplicate.

Teaching a Computer to Fly

This isn’t entirely hypothetical.

In 2023, the Defense Advanced Research Projects Agency (DARPA) demonstrated artificial intelligence controlling a specially modified F-16, designated the X-62A VISTA, during simulated air-to-air combat engagements against human-piloted aircraft.

In May 2024, Secretary of the Air Force Frank Kendall flew aboard the aircraft while AI performed tactical maneuvers without either occupant touching the controls.

By July 2026, the Air Force and DARPA had progressed to testing AI-controlled flight in modified operational F-16s through the VENOM program.

These developments don’t mean AI is ready to replace every combat pilot tomorrow. Flying an aircraft, conducting combat missions, and exercising military judgment are different problems.

Nevertheless, we’ve demonstrated something remarkable. Artificial intelligence can control a high-performance fighter through complex aerial maneuvers in a real-world environment.

Now let’s see it change a hydraulic line.

The Trouble with Replacing a Mechanic

Flying an aircraft is an extraordinarily demanding activity, which I know firsthand from my F-16 incentive flight. But from an engineering standpoint, the pilot operates through a relatively limited number of physical interfaces.

There are flight controls, throttles, switches, displays, communication systems, and sensors. These interfaces are already designed to translate human commands into mechanical, electrical, and now mostly digital actions.

A computer doesn’t need hands to fly a modern aircraft. It needs access to the appropriate control systems and sufficient information to make decisions.

Aircraft maintenance is an entirely different proposition.

Imagine instructing a robot to replace a component buried deep inside an aircraft fuselage.

First, it must remove an access panel secured with fasteners that may have endured years of vibration, corrosion, temperature changes, and abuse from previous maintainers.

Then it must reach through an opening apparently designed for a six-year-old contortionist, locate a component it cannot fully see, disconnect several electrical connectors, loosen fittings without damaging adjacent equipment, and extract the defective part without destroying anything else.

It must then reverse the process, correctly secure everything, and verify the repair.

A human mechanic might accomplish this while balancing on a maintenance stand, holding a flashlight between his teeth, and questioning the ancestry of the engineer who designed the installation.

The robot needs extremely sophisticated vision, tactile sensors, manipulation capabilities, and sufficient intelligence to recognize when something isn’t going according to plan.

And that’s just one maintenance task.

What happens when the replacement component doesn’t fit quite right? When a connector refuses to seat? When a bolt begins to cross-thread? When a fitting moves but the line behind it starts twisting?

An experienced mechanic can often recognize these problems through touch before any visible damage occurs.

Replicating that capability is considerably more complicated than teaching software to manipulate flight controls.

The Aircraft Wasn’t Designed for Robots

This brings us to a fundamental problem.

Modern military aircraft were designed to be maintained by humans.

Granted, some aircraft designers appear to have interpreted that requirement as a personal challenge to see how miserable they could make the experience.

Nevertheless, access panels, component locations, connectors, fasteners, and servicing points generally assume a human being will perform the work.

A robot attempting to maintain an existing fighter must adapt to thousands of engineering decisions made without considering robotic access.

That seems like an unnecessarily expensive way to approach the problem.

The more logical solution would be to design an aircraft specifically for robotic maintenance from the beginning.

Imagine an aircraft whose avionics are packaged into standardized cartridges that robotic equipment can remove and replace.

Electrical connections could be designed for automated engagement. Hydraulic systems could incorporate robotic servicing interfaces. Access panels could use standardized mechanisms rather than a collection of fasteners seemingly selected by throwing darts at a hardware catalog.

The aircraft could automatically diagnose discrepancies and communicate with the maintenance system before landing.

By the time it reaches its parking spot, a replacement assembly could already be waiting.

Instead of building a robot capable of reproducing every movement of a human aircraft mechanic, engineers could design the aircraft so that maintenance tasks require relatively simple, repeatable mechanical actions.

This would be a significant departure from conventional aircraft design.

Unfortunately, it introduces another rather inconvenient requirement.

The Aircraft Still Has to Win the Fight

An aircraft designed for easy robotic maintenance would be a remarkable engineering achievement.

It would also be completely worthless if an enemy fighter could shoot it out of the sky because its designers compromised combat performance to make replacing components easier.

Stealth aircraft provide an excellent example.

Their external shapes, surface treatments, panel arrangements, and structural configurations are carefully engineered to reduce detectability.

Adding convenient robotic access doors, standardized attachment points, and modular equipment compartments could introduce unwanted weight, structural compromises, and radar-signature problems.

And fighter aircraft designers are notoriously protective of weight.

A few additional pounds here and there may seem insignificant until they affect range, acceleration, maneuverability, or weapons capacity.

There is also the issue of survivability.

A conveniently removable component might also be more vulnerable to combat damage. A standardized connection designed for robotic convenience might introduce another potential failure point.

Engineers would have to reconcile two competing priorities: making the aircraft exceptionally capable in combat and making it exceptionally easy for machines to maintain.

Those objectives won’t always agree.

The resulting aircraft might actually be more expensive to manufacture than a conventional design.

Which raises another uncomfortable question: would the savings in maintenance costs justify the compromises in aircraft performance?

That’s not a decision anyone should make based solely on how impressively a robot can turn a wrench.

The Economics of Replacing Maintainers

There is another reason I suspect fighter pilots may face artificial competition before aircraft mechanics.

Money.

A sophisticated maintenance robot might cost hundreds of thousands of dollars, or considerably more, depending on its capabilities.

To justify that investment, it would need to perform enough useful work to offset acquisition, integration, operating, and support costs.

A human aircraft mechanic can perform an astonishing variety of tasks.

One day might include servicing hydraulic systems, removing panels, inspecting flight controls, troubleshooting electrical discrepancies, assisting with an engine change, and recovering aircraft.

A robot designed to perform one or two of those tasks might spend most of its time sitting idle.

Certainly, specialized robots already have practical applications in aviation maintenance.

Automated inspection equipment can examine aircraft surfaces. Robotic systems can assist with repetitive drilling, coating removal, and other labor-intensive operations. AI can help analyze inspection imagery and maintenance data.

The Air Force is already experimenting with robotic inspection systems employing ground robots and drones.

But there’s an enormous difference between a machine that detects damage and a machine that repairs it.

The former is increasingly practical.

The latter may require an extraordinarily expensive combination of robotics, sensors, artificial intelligence, and specialized tools.

And if the machine gets confused, somebody still has to troubleshoot the robot or repair the damage it created. CNC machine operators have felt this pain.

I can already imagine the maintenance forms.

Discrepancy: Robotic maintenance unit failed to remove aircraft access panel.

Corrective action: Human maintainer removed access panel, repaired damaged robot manipulator, and returned both to service.

Somewhere, a crew chief is shaking his head.

The Problem Gets Worse on Deployment

It’s relatively easy to envision robotic maintenance equipment operating inside a climate-controlled depot facility.

Aircraft are positioned precisely. Components are accessible. Specialized machinery can remain permanently installed. The same operations may be repeated hundreds or thousands of times.

Now move the operation to an austere deployment location.

The maintenance equipment must function in extreme heat, freezing temperatures, blowing sand, rain, snow, inadequate lighting, and less-than-ideal working conditions.

It must tolerate uneven surfaces, equipment shortages, transportation damage, and the occasional unexpected encounter with a vehicle operated by an airman who probably shouldn’t be driving anything more complicated than a shopping cart.

And what happens when the robot breaks?

Do we deploy robotic maintenance technicians to repair the robotic aircraft maintainers?

And who repairs those robots?

At some point, the Air Force may discover that replacing a human mechanic requires an entire collection of machines, support equipment, software specialists, and repair personnel.

That doesn’t necessarily make automation uneconomical. But it certainly complicates the calculation.

Perhaps We’re Designing the Wrong Aircraft

There is another possibility worth considering.

Maybe the future isn’t a conventional fighter maintained by sophisticated humanoid robots.

Maybe it’s an entirely different aircraft maintenance philosophy.

Autonomous combat aircraft could be designed around modular replacement rather than extensive field repairs.

Instead of troubleshooting complicated internal failures on the flight line, robotic systems might simply exchange entire assemblies.

Defective components would return to centralized facilities where specialized machinery and human technicians could repair them.

Some relatively inexpensive autonomous aircraft might even be designed for limited service lives, making extensive repairs economically questionable.

This could fundamentally change the organization of military aircraft maintenance.

But there is a critical distinction between designing an aircraft that doesn’t need a pilot and designing one that doesn’t need a mechanic.

Eliminating the cockpit removes a human requirement from the aircraft.

Eliminating human maintenance requires rethinking much of the aircraft’s physical design, its components, its support equipment, and the entire maintenance infrastructure.

That’s a considerably larger engineering undertaking.

And throughout that process, the aircraft must remain capable of defeating its adversaries.

An exquisitely maintainable fighter that consistently loses aerial engagements is not a technological triumph. It’s an exceptionally efficient way to replenish the enemy’s victory tally.

The Last Man on the Flight Line

I have little doubt that artificial intelligence and robotics will transform military aircraft maintenance in some fashion.

We will see increasingly sophisticated diagnostics, automated inspections, robotic servicing equipment, and machines capable of performing repetitive maintenance operations with remarkable precision.

Some traditional maintenance duties may eventually disappear.

Others will change substantially.

But replacing the entire range of physical skills, mechanical judgment, experience, and adaptability possessed by a competent aircraft maintainer is a much more formidable challenge than many people appreciate.

A fighter pilot operates an enormously sophisticated machine.

An aircraft maintainer must understand, inspect, troubleshoot, repair, and occasionally persuade that machine to cooperate.

Those are fundamentally different problems.

And while artificial intelligence may eventually master both, the physical world has a particularly unpleasant way of complicating otherwise brilliant software.

So, for the moment, I offer my congratulations to the pilot community.

After decades of technological progress, we’ve finally developed computers sophisticated enough to threaten your jobs, especially those in the fighter world.

Fortunately, those same computers still can’t reliably reach around a hydraulic line, disconnect a stubborn cannon plug, or remove a seized fastener without creating three additional maintenance discrepancies.

Perhaps someday they will.

Until then, we’ll be over here at the toolbox, still smelling of sweat and jet fuel.

And when your million-dollar maintenance robot breaks, feel free to call an aircraft mechanic. We can always squirt a little RTV into a stripped nutplate to get one more critical combat flight in the air.

_____________________________

Dave Chamberlin runs a consulting and training company and brings more than 40 years of civilian and military aviation experience to his work. He retired as a Chief Master Sergeant after 38 years as an aircraft crew chief in the U.S. Air Force and Air National Guard, and has also worked in technical, instructor, consultant, and leadership roles. He holds an FAA Airframe and Powerplant license and a master’s degree in aeronautical science, and his writing often focuses on military issues, especially those affecting aircraft maintenance personnel.

As the Voice of the Veteran Community, The Havok Journal seeks to publish a variety of perspectives on a number of sensitive subjects. Unless specifically noted otherwise, nothing we publish is an official point of view of The Havok Journal or any part of the U.S. government.