Blitzkrieg Redux: The Coming Warbot Revolution
In May 1940, the Allied and German Armies squared off in what was expected to be an extended campaign for the conquest of France. Six weeks later, the victorious German Army marched down the Champs-Elysees in Paris. How was it that the Germans, with fewer tanks, fewer trucks, fewer troops, less artillery and access to roughly equivalent technologies, managed to accomplish such a remarkable feat? While leadership, luck, and a host of other factors were at play, the decisive factor was the remarkable way in which a few German inter-war military thinkers envisioned and developed a new way of warfare, known to the Allies as the blitzkrieg. German doctrine successfully integrated current technologies in aircraft, radios, and tanks into a coherent and integrated way of fighting and then applied it to great effect.The result was amplified because the Germans fought an enemy that in many cases failed to account for the possibilities enabled by the new combination of these technologies.
We are now on the cusp of a similar revolution in warfare with the opportunity to integrate several current and near term technologies into our concept of how we will conduct military operations in the not-to-distant future. The winner of the next conflict will not likely be determined primarily by the state of their technologies, but by how well a nation’s military thinkers conceptualize future warfare in an integrated manner and then apply robotic systems, or warbots, appropriately to our way of fighting. For purposes of this discussion, warbots can be defined as robotic combat systems that can detect, identify, and apply lethal force to enemy combatants within prescribed parameters and without immediate human intervention. Using the historical lens of the blitzkrieg, we will examine two key trends that can help inform our concept of future warfare and our ability to wage it. They include: the rise of lethal warbots as primary combatants and adapting current leadership methods to a future era of manned-unmanned, or Centaur, teaming.
THE RISE OF LETHAL WARBOTS AS PRIMARY COMBATANTS
French armored doctrine used tanks to support slow moving infantry armies. While generally armed with heavier guns and having thicker armor than most German tanks, they were also slower, rarely had radios to enable them to fight effectively in groups or to exploit rapidly changing battlefield conditions, and often had ineffectual one man turrets.[i] In contrast, the Germans viewed their tanks as primary fighting vehicles that would spearhead deep attacks into the enemy’s rear as part of a combined arms team.[ii] The Germans consequently put a premium on balancing armor, firepower, maneuverability, and reliability. They also equipped most combat vehicles with radios to facilitate both effective maneuvering of large formations and rapid coordination with air assets.[iii]
Most militaries continue to look at warbots as support weapons that can conduct reconnaissance, selective strike, and logistical or other supporting tasks. Many military leaders are most comfortable with warbots in these limited rolesbecause it is easier to keep humans in the loop and therefore retain a greater feeling of control. There is comparatively little discussion of incorporating warbots in significant, and in certain scenarios, primary combat roles into future combat doctrine. To do so, and to take advantage of the full capabilities of warbots, would require acceptance of a greater degree of autonomy, akin to the Army concept of Mission Command, than the US military seems to be comfortable with at this point. The U.S. military continues to invest in manned combat systems, yet warbots offer tremendous potential advantages as primary combatants. They are simply more capable, cheaper, and offer less risk to humans than manned equivalents in many, if not most, combat situations.
Robotic systems are able to engage enemies and respond to threats at far higher speeds than humans.The AH-64D Apache attack helicopter’s Longbow fire control radar already “automatically searches, detects, locates, classifies, and prioritizes multiple moving and stationary targets on land, air, and water in all weather and battlefield conditions.”[iv] Adding the ability to engage within human “on the loop” specified parameters offers the advantage of getting the first, and likely lethal, shot off faster than a human “in the loop” configuration. While the Israeli Trophy Active Protection System (APS) is a defensive system, it further demonstrates how faster reaction times could be achieved in an offensive system. It is a “fully automated” active point defense for vehicles thatresponds without human intervention to rapidly detect and neutralize incoming rockets and missiles with both “shotgun like blast[s] of pellets” as well as jamming.[v] Inserting a human into this decision loop would degrade the system and put it at a disadvantage due the necessity of ultra fast reaction times to counter an incoming missile.
Another limitation of humans versus warbots is that the humans themselves are often the limiting performance factor. A pilot will black out at approximately nine Gs, but an appropriately designed airframe can handle much more.[vi] While robotic systems have unique vulnerabilities, to electro-magnetic pulses for example, humans are generally far more vulnerable than vehicles to a host of lethal and non-lethal factors. These include kinetic, nuclear, chemical, and biological weapons, as well as the combat degrading effects of sleep deprivation, temperature extremes, and emotional factors such as anger, hatred, fear, and even the desire for self preservation. The use of warbots also has the potential to stretch the combat power of defense budgets. Humans are not expendable, and much of the weight and cost associated with modern aircraft, ships, and vehicles is due to the necessity of keeping them alive to operate the combat systems. But what if warbots were both effective and relatively expendable?
One B-21 is projected to cost $550 million, not including the personnel costs associated with the lifecycle costs of the pilots needed to fly it, the costs of the Combat Search and Rescue infrastructure needed to recover the pilots should they be shot down, and the other additional incremental costs of a human piloted system. A similar situation exists for land combat vehicles. Modern tanks are limited by the human crews and the necessity to protect them. An armored vehicle without onboard humans does not need the size or safety features required to keep them alive. It can be made smaller, lighter, and potentially disposable (in relative terms). This could also reduce the Army’s manpower requirements as warbots don’t draw pay, need healthcare, require rehabilitation when injured, or collect retirement.
While warbots will never fully replace manned ships, vehicles and aircraft in all situations, they are too good to pass up as they offer increased effectiveness, reduced cost, and less risk to humans.
ADAPTING CURRENT LEADERSHIP METHODS TO CENTAUR TEAMING
Leadership in a military unit serves as botha combat multiplier and a unifier of all the diverse elements of combat power.[vii] Advances in robotics will require adapting how we lead military forces consisting ofhumans and employing increasingly large numbers of warbots. We are often concerned by a perceived lack of control over warbots, and while this is a legitimate concern, it cannot be one to which we retreat intellectually. To maximize the utility of large numbers of warbots, leaders will have to adapt their comfort zones to accept a certain degree of decentralization in their activities. Just as the German Army’s degree of comfort with decentralized execution was a decisive advantage in the French Campaign, the U.S. Army’s concept of mission command gives it a decisive advantage on the battlefield without even the expectation of perfect control over every soldier and vehicle. While absolute autonomy is correctly viewed as unacceptable, if we are to see the full benefits of warbots, we already have the mission command framework and can adapt it to the new forms of interaction that will emerge between manned and unmanned systems.
In manned-unmanned teams, or Centaur teams,[viii] artificial intelligence does well in areas that require exceptionally fast and accurate calculations such as missile intercept solutions, and with long-term repetitive actions, such as extended surveillance of a fixed location. Artificial intelligence beat the world chess champion in 1997[ix] and has advanced to beating the world Go champion.[x] Yet these systems often struggle with tasks that require a combination of judgement and calculation capability. The solution may be found in the appropriate level of Centaur teaming, in which humans and computers are employed together with a resultant improvement in the abilities of both. As an example, Centaur teams play better chess than either humans or computers can independently.[xi] While Go and chess are fairly bounded games with specific rules--in other words, an activity at which artificial intelligence is likely to do quite well--warfare is far less so and infinitely more complex. We must ask ourselves where the sweet spot is between where we can enable AI more fully, and where humans can and should manage components of an engagement to maximize the capabilities of both. Interestingly, a clue may emerge as Google DeepMind works with Blizzard Entertainment to use its hugely popular game StarCraft II as a research platform for artificial intelligence.[xii] The complex, real-time game involves resource allocation, combat, and the defeat of an enemy while managing an army that is enabled by relatively primitive artificial intelligence. This type of gaming technology and interfaces can offer useful tools to enable future commanders control large numbers of semi-autonomous warbots on a complex and changing battlefield. Perhaps multitasking skill in this environment may even become a criterion for selection of future leaders of tactical warbot forces. Ultimately, the conceptual jump from the army’s current mission command philosophy with humans to mission command of Centaur teams is no cognitive leap across a canyon..
A CAVEAT AND CONCLUSION
While warbots offer promising possibilities, the nature of war itself has not changed, nor is it likely to. War remains, in Clausewitzian terms, “an act of force to compel the enemy to do our will.”[xiii] It is enabled by weapons, but it is ultimately won or lost by humans and their decisions. Likewise, securing a better peace following a war (consolidation of gains, stability operations, transitional military government, etc.) will remain nearly exclusively human in their application. For all their lethality in war, warbots lack the human abilities to empathize, dispense justice, interact with and exercise authority over humans, mitigate suffering, reason morally, and interact with humans in a meaningful way.
Americans often excel at innovation and the U.S. military can decisively seize the warbot initiative by conducting the same bold experimentation as seen in the pre-war Louisiana maneuvers. Just as these maneuvers tested concepts that helped lay the groundwork for success in World War Two, we have the opportunity to further develop our doctrine to incorporate both warbots and Centaur teaming, or “Centaur Warfare,” before our our future enemies do. Warbots are coming, the only real question is whether we will model our approach after pre-war France or Germany.
For further reading on what future ground combat might look like, please see “Warbot: The Death of Homer,” published by the Atlantic Council’s Art of the Future Project.
Brian M. Michelson is a U.S. Army Officer and currently serves as a Chief of Staff of the Army Senior Fellow at the Atlantic Council. His previous writing has appeared in Military Review, Small Wars Journal, and the Institute of Land Warfare. The views expressed in this article are those of the author and do not necessarily reflect the official policy or position of the United States Army War College, the United States Army, the Department of Defense, or the United States Government.
[i] Battle of France. Wikipedia, https://en.wikipedia.org/wiki/Battle_of_France#Analysis. Accessed February 23, 2017.
[iv] Longbow™ Fire Control Radar and RF Longbow HELLFIRE™ Fire-and-Forget Missile for the Apache. Lockheed Martin / Northrop Grumman, 2007, http://www.northropgrumman.com/Capabilities/LONGBOWFireControlRadar/Documents/longbow.pdf. Accessed February 17, 2017.
[v] Kyle Mizokami. Army and Marines To Test Israeli-Built Missile-Killing Tank Defense System. April 22, 2016, Popular Mechanics, http://www.popularmechanics.com/military/weapons/a20503/army-marines-killer-tank-defense-system/. Accessed February 17, 2017.
[vi] Peter Tyson. All About G Forces. November 1, 2007, NOVA, http://www.pbs.org/wgbh/nova/space/gravity-forces.html. Accessed February 17, 2017.
[vii] ADP 6-22, Army Leadership. Department of the Army, August 1, 2012, p. 1.
[viii] Sydney J. Freedburg, Jr. Centaur Army: Bob Work, Robotics, & The Third Offset Strategy. November 9, 2015, Breaking Defense, http://breakingdefense.com/2015/11/centaur-army-bob-work-robotics-the-third-offset-strategy/. Accessed February 20, 2017.
[ix] Almanac: Kasparov vs. Deep Blue. May 3, 2015. CBS News, http://www.cbsnews.com/news/almanac-kasparov-vs-deep-blue/. Accessed February 20, 2017.
[x] Jon Russell. Google AI beats Go world champion again to complete historic 4-1 series victory. March 15, 2016, TechCrunch, https://techcrunch.com/2016/03/15/google-ai-beats-go-world-champion-again-to-complete-historic-4-1-series-victory/. Accessed February 20, 2017.
[xi] Mike Cassidy, Centaur Chess Shows Power of Teaming Human and Machine. December 30, 2014, Huffington Post, http://www.huffingtonpost.com/mike-cassidy/centaur-chess-shows-power_b_6383606.html. Accessed February 22, 2017.
[xii] Oriol Vinyals. DeepMind and Blizzard to release StarCraft II as an AI research environment. November 4, 2016, Deepmind, https://deepmind.com/blog/deepmind-and-blizzard-release-starcraft-ii-ai-research-environment/. Accessed February 20, 2017.
[xiii] Carl von Clausewitz. “On War.” Edited and translated by Michael Eliot Howard and Peter Paret. Princeton University Press, 1976, p. 75