So, there I was. In the Combined Air and Space Operations Center (CAOC) at night. Inside, it always looked the same, regardless of the time of day and intensity of the harsh desert sun outside. The room was cold and dark, lit by flickering monitors and large wall screens. Only the clacking of keyboards, humming servers, and murmured conversations disturbed the quiet.
I was the JSTARS Master Air Attack Plan Liaison Officer, or MAAP LNO, the person responsible for planning where our aircraft went, making sure they got there with fuel and a plan to get on station, and then handing the night off to the mission LNO.
That is when the email came in: a request to provide ground-moving target indicator (GMTI) support for a dynamic operation that had just begun. The phone rang soon after. “Can your guys look at that spot tonight? High priority, this op just popped up.”
I sighed as I scrolled the map to the planned JSTARS orbit for that night. I punched in the coordinates for the new location. “Hang on, let me look,” I said, with what probably sounded like disdain. Not at my lack of desire to support the mission, but for the work I knew was coming. Eyeballing it, I answered, “Yeah, we can do it.”
I hung up the phone and pushed aside my dry chicken and rice from the dining facility. Now I had to determine where the jet would orbit to keep line of sight on the high-priority target while covering our other jobs and decide which lower-priority requests to drop for the night. The new plan needed to go to the mission planning team, and I still had to check if this would affect our tanker needs, timing, or location later in the sortie.
The average American at the time might have assumed there was a faster, more sophisticated way to do this. There was none. Most of it had to be done by holding a transparency sheet over the screen, making measurements, entering pre-drawn shape files, and doing math with a highly sophisticated (read: unwieldy) gonkulator. All this happened while bouncing back and forth between secure and unsecure systems that only sometimes did what I needed them to do.
And I was a political science major.
What Changed
That was a decade ago. In just half my career, the tools have changed almost beyond recognition. The pace of change is no longer linear. What once felt like a stretch goal in that CAOC—some faster way to model an orbit or rework a tanker plan—is now baseline capability. What is coming makes that baseline look quaint.
Since then, I have worked in jobs where that future is planned for and built. I helped divest old capabilities and evaluate what was next for C2 on the Air Staff. I assisted a Middle East partner in modernizing their force to help meet American security cooperation goals. I assessed threats and opportunities of AI, dual-use technology, and autonomous systems while also designing wargaming scenarios at the Defense Threat Reduction Agency to stress-test how institutions might respond to those systems.
None of this fully prepared me for how bluntly David Petraeus and Isaac Flanagan put it in a recent piece for Foreign Affairs titled The Autonomous Battlefield: And Why the U.S. Military Isn’t Ready for It. Their argument cuts to the heart of something I’d been turning over for a while. The nature of C2 itself is changing, not just the tools we use to do it.
Their argument is blunt. Militaries that insist on human control at the tempo of human decision-making will lose to adversaries that do not. The side that waits for a person to approve of every action before acting will be outpaced by the side that does not. This is already happening in Ukraine. It will define the next era of warfare.
I will be honest: I find myself in an uncomfortable position. I agree with nearly all of it. Not reluctantly, exactly, more like the way you agree with a diagnosis you already suspected but were not ready to hear confirmed. This is not a future I fear, though. The Air Battle Manager (ABM) career field has learned to move on from old ways when better ones emerged. But this future has arrived faster than the institutions built to manage it. That gap—between what technology can do and what doctrine prescribes—is where the real risk lives.
But this is not an obituary for the Air Battle Manager. It is, if anything, the opposite.
The Adaptation
Petraeus and Flanagan are not saying humans should disappear from battlefield operations. They argue something more precise and more unsettling. “Execution — the sensing, targeting, movement directions, timing, and striking — will shift to algorithmically piloted machines that humans program but do not control moment to moment.” The human role does not vanish; it moves upstream.
For an Air Battle Manager, that distinction is enormous. The career field has always managed the air picture in real time, reading the environment, redirecting assets, and managing the net when things go sideways. That is the job: bring order from chaos. But if Petraeus and Flanagan are right, the future job looks less like managing execution and more like designing the conditions so execution can continue without you. Not tactical control, but mission architecture.
The shift Petraeus and Flanagan describe, from controlling systems in battle to preprogramming them, requires an innovative approach. Commanders must “translate their intent into terms precise enough for machines to execute.” They specify not just objectives, but actions permitted, actions forbidden, boundaries, abort criteria, and what to do when conditions change. For those of us trained to decide in real time — from the seat, with the full picture in front of us — this is a fundamentally different cognitive task. In some ways, it is harder and demands a different rigor. The ambiguity that an experienced ABM can resolve in seconds at the scope must now be resolved before the mission launches. It must be baked into architecture because a human may not be available when it matters or may be too late to act if they are.
Petraeus and Flanagan call this the “command-design problem.” It is the central challenge of this era. The winner, they argue, “will not be the side with the most drones but the side that best solves the command-design problem.” The challenge is to design command for systems that execute at machine speed, out of contact, within preset boundaries.
That is an operational C2 problem.
It has always been an operational C2 problem. The ABM career field has spent decades thinking about how to structure the air picture so execution can continue when the net degrades, communications are jammed, or the situation changes faster than the plan. The autonomous era does not invent that problem. It amplifies it by orders of magnitude and removes the margin for error that human judgment in the loop once provided.
The field is well-positioned to own this problem. But only if it moves deliberately toward it rather than waiting for the doctrine to catch up.
Petraeus and Flanagan are clear about where the institution stands now. “No joint doctrine for autonomous formations yet exists. No major command has been tasked with developing one.” The doctrinal vacuum is real. The education gap may be deeper. Military institutions “will need to mint a new generation of commanders adept at programming algorithms, or directing programmers, to execute operational objectives.” They must treat “software engineers, data scientists, and electronic warfare specialists as essential staff.” For the ABM career field, that last part should feel familiar in spirit, if not in practice. The field has always worked at the intersection of technology and operational command. The step from managing the systems others built to shaping requirements and constraints is not as large as it appears. But it requires intention. It demands that the career field join the design conversation early, not as a user accepting what the acquisition system delivers, but as the community that knows best what command at machine speed requires.
None of this is comfortable. Some of it is genuinely unresolved. Petraeus and Flanagan acknowledge that “speed without adequate governance can yield errors and unintended escalation.” Autonomous systems engaging at machine speed can trigger similar responses, causing “an escalated exchange that no human commander intended or authorized.” The accountability questions that follow do not yet have clear answers. They may not for some time. But the career field responsible for the air picture—knowing what is where, what it is doing, and what happens if it does something else—needs to be in that conversation.
Conclusion
I said at the outset that this is not an obituary. I meant it.
The Air Battle Manager career field was built to solve hard problems at the intersection of technology, doctrine, and operational command and control. It has done that through every iteration of the fight since the career field’s inception. From the radar scopes of Control and Reporting Centers to the consoles of AWACS and JSTARS aircraft, to operations center floors, all over the world. The tools changed. The problem did not. Provide the commander with a clear air picture. Find, fix, and track targets, facilitate engagement, and assess. Make the right call when the situation changes faster than the plan. Bring order. Adapt and repeat.
None of that is going away in the autonomous era. It is getting harder, faster, and more consequential. The question is not whether the ABM career field has a role in what is coming. It does, and it is central. The question is whether the career field gets ahead of the adaptation or waits to be told what it looks like.
I am confident it will get ahead of it. That confidence is built on twenty years of watching this community figure out hard things under pressure, without enough resources, faster than anyone thought possible.
But I will leave you with the question I have been sitting with since I read Petraeus and Flanagan, the one I do not think has a clean answer yet, and the one the career field needs to be wrestling with aloud:
If humans cannot keep up with the tempo of war, what exactly are we choosing to stay in control of?
Your radios.
Lt Col Jason “JB” Baker is a retiring Master Air Battle Manager instructor and evaluator veteran pursuing a career in defense policy and AI governance. He has also served at Headquarters Air Force, the U.S. Military Training Mission to Saudi Arabia, and the Defense Threat Reduction Agency.
The views expressed are those of the author and do not reflect the official policy or position of the U.S. Air Force, Department of Defense, or U.S. government.
Photo by Mael BALLAND on Pexels
Leave a Reply