The Future of Warfare Will Continue to Be Human


The dustbin of history is full of predictions of technological innovations that would surely change the nature of warfare. Machine guns, tanks, and modern artillery were all supposed to change warfare radically, yet within a short time of their fielding, less tech-savvy forces were able to mitigate these advances through tactical innovation. Frequently, technological have-nots went on to seize victory against more technologically advanced adversaries.

Will the Pentagon’s current pursuit of tech overmatch, with artificial intelligence (AI) as its crown jewel, pay off? Or, will resources devoted to AI risk crowding out investment in the enduring human determinants of battlefield victory or defeat?

Will AI Be Revolutionary?

Conflict in the future will take place in a battlespace that is shaped by artificial intelligence and other new technologies. Some more bullish observers suggest that this change will be revolutionary, that the leader in AI will rule the world, and that soon war may not even have humans involved. Though faith in the transformative potential of technology is tempting, hope that robots will replace humans in war is misplaced. Since the dawn of the industrial age, technological advantage has been as predictive of victory as a coin toss.  Instead, human innovation has overcome technical disadvantages to deliver victory to those who most quickly exploit the battlefield.

If force employment, doctrine, and tactical decisions made by humans trump technology in determining victory, too much focus on artificial intelligence could create blind spots that future adversaries will exploit. Though AI and other technology will be important in future conflicts, the best bet for future victory is developing the tactical, operational, and strategic innovators who will leverage battlespace conditions to achieve victory, regardless of the tech balance. Human intelligence and creativity will win the next war, not technology.

Instead of seeking to replace it, virtual training environments should be used today to help develop the innovative potential that will deliver victory tomorrow. Michael Horowitz and Casey Mahoney reflected on the state of the conversation about AI over a year ago. A consensus seems to have grown, at least in the pages of War on the Rocks, that AI and other emerging technologies are changing the character of warfare, if not its nature.  Though AI will be a powerful tool, it’s unlikely that the world will be ruled by those who develop the most advanced AI-based weapon systems. First, the glacially paced military change of recent memory suggests that battlefields fifteen years hence will likely look more familiar than some may think. Second, military analysts have been erroneously forecasting revolutionary change just around the corner for at least four decades. And third, research demonstrates that technological overmatch does not significantly increase the likelihood of military victory.

The battlespace of 2035 is now fifteen years distant. I joined the Air Force a little more than fifteen years ago, and those years have been mainly characterized by slow evolution rather than revolutionary change. The 1960’s era radar system I first trained on is still our primary ground mobile radar system. A next generation replacement has been in the works since I joined, but after fifteen years of anticipation, the contract was recently  cancelled and the Air Force is returning to the drawing board. Meanwhile, new generations of airmen are training on the same sixty-year-old radar system. Air operations have experienced evolutionary change since I trained as an air battle manager a few years later, and the systems we use for battle management and command and control still resemble home desktops in the 1990s more than the machines in our pockets today. If the last fifteen years are any indication of the next, we will continue to see incremental, evolutionary increases in capability even as we approach 2035 and the 50th year of the information age of warfare.

While it’s possible that AI will introduce revolutionary change, that seems unlikely. This is where the work of Stephen Biddle and others suggest some need for expectation management: “A mathematical property of exponential growth is that it always displays a rapidly growing rate of change with a sharp upturn at the end of the period, whatever period one considers.” Simply put, exponential curves produce a mathematical optical illusion that causes the present moment to appear to be one of near vertical growth.

From the development of tanks and machine guns in World War I to increases in artillery range and munition lethality, to the precision guided weapons of the information age, Biddle describes increases in battlefield lethality as growing exponentially since the beginning of the industrial age. During that entire period, military analysts routinely concluded that their current period was on the cusp of a revolution in military affairs. An observer of exponential growth in battlefield lethality in 1950 would be just as likely to predict an immanent revolutionary transformation as one in 2000… or 2020. The curve looks identical at all scales. This suggests that exponential advances in lethality since the turn of the previous century are more evolutionary than revolutionary in any given period, though they always appear on the cusp of revolution.

This leads to Biddle’s second, and perhaps more surprising finding: Over the last century tech overmatch has been about as predictive of victory as a coin toss. Biddle’s multimode analysis has shown that of 16 wars between 1956 and 1992 for which data was available, the technologically superior side won eight times. In his 2006 book Military Power: Explaining Victory and Defeat in Modern Battle, Biddle argues that massive increases in lethality associated with the machine gun, airpower, tanks, and other 20th-century technological advances did indeed change the way modern militaries fight.

However, the change is primarily due to tactics developed on the battlefield to minimize exposure of friendly forces to those advances in weapon systems. Simply put, when advances in artillery and airpower rendered massed forces virtually suicidal, warfare didn’t simply become an exchange of vaporizing massed forces. Instead, commanders dispersed their forces and innovated tactics that enabled the accomplishment of military objectives while minimizing exposure to increasingly lethal weapons. Though the battlefield learning curve is steep and costly, those who innovate tactically go on to victory as often as those who possess the most advanced battlefield technology.

Harnessing Virtual Wartime Selection Pressures

Technology does help win wars, and AI has important military potential. However, if technological advances in lethality are overcome by tactical innovation as often as not, then tech overmatch is a “nice to have,” not a “need to have.” To avoid the worst of the “harsh wartime selection pressures” that are waiting for the unprepared, innovation in the use of force should be prioritized at least as highly as innovation in weapon systems. To do this, two things are needed: first, greater investment in virtual training capability and capacity, and second, reducing operations tempos to enable greater focus on readiness for the future fight.  Simply put, this means significantly reducing U.S. share of force commitments in Syria, Iraq, and Afghanistan so that U.S. personnel can focus on training for the next war — likely against a major power like China or Russia — instead of fighting the last.

The military should invest in a more robust virtual training infrastructure that can confront soldiers, sailors, airmen, marines, and space professionals with the joint tactical dilemmas they will potentially face in peer competition in the future. Currently, large exercises such as RED FLAG face serious challenges of scale and complexity if the goal is to simulate wartime pressures. These exercises tend to simulate a single isolated mission that is not well integrated into a wider operation. There is often limited or no play from command elements and joint partners, which renders the mission hermetically sealed from the dynamic forces of a wider joint operation. There are also significant challenges in realistically integrating multidomain operations into a legacy exercise construct. The result is great training, but it doesn’t increase joint integration or encourage rapid joint innovation.

Fortunately, this is one of the areas where AI can be applied in the near term to help warfighters without suffering the costs that modern lethality can inflict. Many of the building blocks are in place already. The Air Force has been using distributed virtual training for years and today many units have local access to Distributed Mission Operations training that allows local operators to work with remote instructors. The Distributed Mission Operations Center at Kirtland Air Force Base provides mock-ups of  a number of battle management platforms in a virtual training space and enables complex integrated training scenarios. Virtual Flag exercises routinely integrate joint forces in virtual environments, and efforts like Pilot Training Next and other “gamification” of training efforts demonstrate the potential on the horizon to integrate geographically separated units into innovative training environments without most of the logistical hurdles of live training operations. Most promisingly, Live, Virtual, Constructive training offers the potential to augment live exercises with virtual and constructive elements, significantly reducing the realism tradeoffs that today define most live training. Paired with AI, there is near-term potential for warfighters to integrate with simulated joint partners, as well as adversaries, in order to enable far more advanced and realistic training without radically increasing the bandwidth of scenario designers and white force players.

Second, difficult choices are needed to enable greater focus on innovation in force employment.  Current operations tempos place very heavy burdens on readying warfighters for today rather than tomorrow. In his testimony to the Senate Armed Services Committee on the 2018 National Defense Strategy, Elbridge Colby testified that the “U.S. armed forces should become, as in most of the Cold War, primarily a training and readiness-oriented force prepared for war against a near-peer opponent, and not, as in the post-Cold War period, a military largely focused on operations in the Middle East.” In recent Defense posture hearings, House Armed Services Committee Chair Adam Smith and others have called for hard choices in order to support National Defense Strategy. One of the costs of not making these choices is that operations tempos today leave little bandwidth for serious dedication to the questions of how we will fight tomorrow. Meanwhile, Pentagon leaders are enthusiastically encouraging warfighters to take risk, “fail fast,” “fail productively,” and experiment as an iterative learning process.

While calls for more training are reasonable, it poses a unique challenge to current warfighters who are caught between formal training requirements and the pull of real-world operations, both of which discourage significant risk taking. Instead, requirements, acquisitions, and other staff officers are trying to buy the gear of the future for warfighters who have not yet figured out whether they need it, or how they would use it. It is a mistake to focus the creative efforts of talented military members primarily on acquiring new technology at the expense of innovations in force employment, and it is hard to see how that will be possible as long as warfighters are primarily dedicated to the operations of today. As we have seen over the last few years, pivoting away from operations in the Middle East toward anywhere else is easier said than done. But, rapidly transitioning the armed forces into an artificially intelligent, partially autonomous future fighting force isn’t easy either, and it’s hard to see why the former should be unrealistic while the latter is in our reach.


The character of war continues to evolve, and there is no doubt that AI will significantly contribute to elements of that evolution for the foreseeable future. However, there are risks to an overestimation of the rate of technological change and the role of advanced tech in future victory. In spite of many exciting technological developments, history and data suggest that we are probably in an evolutionary rather than revolutionary period. Furthermore, at least since the beginning of the modern age technological overmatch has been neither necessary nor sufficient for victory, and is about as predictive as a coin toss.

In its understandable zeal for technological advantage, the military should remember enduring truths about warfare. The nuance of force employment is key to any battle, and an over-emphasis on technology risks allowing blind spots that our future adversaries will be sure to exploit. Though we should certainly strive for cutting edge technology, if that pursuit comes at the expense of force employment investments, we may be unknowingly entrusting future success to little more than a coin toss.



Maj. Peter L. Hickman has a PhD from Arizona State University in International Relations and a Master of Military Operational Art and Science in Joint Warfare. He is currently a Defense Legislative Fellow for a member of the House Armed Services Committee. Prior to this position, he worked as a Requirements Manager on Air Combat Command HQ staff and the Chief of Weapons and Tactics at the 225th Air Defense Squadron. The perspectives presented through War on the Rocks are his own and do not necessarily reflect those of the Department of Defense, USAF or any other office or organization.

Image: U.S. Army