The Race to Consolidate Power and Stave Off Disaster
In a recent Foreign Affairs article, Jacquelyn Schneider and Julia Macdonald argue, based on their research interviewing U.S. ground troops, that troops prefer close air support from inhabited (“manned”) rather than uninhabited (“unmanned”) aircraft, or drones. After citing the limitations of today’s drones and anti-drone cultural attitudes among troops, they go on to say that “building better drones will not solve this problem” and that “policymakers should reexamine their apparent commitment to an unmanned future.”
Schneider and Macdonald’s research highlights important limitations of today’s drones. They are wrong, however, to conclude that the United States should reconsider its commitment to robotic technology. Quite the opposite: building better drones can solve many of the concerns they raise and should be a priority for future force development.
THE SODA STRAW
Schneider and Macdonald identify two main issues with how troops think about drones. The first is an “engineering problem”—today’s drones are limited by what is known as the “soda straw” view that their pilots have of the battlefield. Drone pilots can see high-definition video of events on the ground but lack the wider field of vision that would help them contextualize what they are seeing. This can particularly be a problem when aircraft are supporting ground troops in combat, a sometimes confusing situation where mistakes can lead to fratricide. By contrast, a human physically sitting in a cockpit can absorb information about the battlefield far more rapidly, particularly if sitting in an aircraft optimized for close air support, such as the A-10. There is no easy way to transmit enough data back to a remote pilot to re-create that same degree of situational awareness.
The current paradigm for remotely controlling aircraft, however, is not the only possible one. For close air support, a better approach may be to actually place control of the air strike directly in the hands of the tactical air controller on the ground, who will have much greater situational awareness than a drone pilot miles away from the battlefield. The Persistent Close Air Support program created by the Defense Advanced Research Projects Agency (DARPA) began to move in this direction, allowing ground controllers to see the aircraft’s video feed and direct air strikes via a tablet. It would be interesting to see whether ground controllers’ objections to uninhabited aircraft change if they are directing the aircraft themselves.
SKIN IN THE GAME
The second issue that Schneider and Macdonald raise is a deeper and more fundamental complaint about uninhabited systems: drone pilots’ reduced exposure to risk. Ground troops directing air strikes expressed concern that drone pilots were not physically “in the fight” and therefore had less skin in the game. This objection is more intrinsic to the nature of uninhabited systems and cannot be as easily overcome by a change in the command-and-control paradigm. Schneider and Macdonald conclude that in situations in which troops “are in direct physical contact with the enemy,” they will continue to be “reluctant to delegate decisions to machines.” This presumption bears further scrutiny.
There are already cases, such as those of bomb disposal robots, in which troops accept that reduced risk is precisely the point of using robotic systems. This suggests a different perspective from the one proposed by Schneider and Macdonald: that troops actually embrace these systems if they help reduce their own exposure to harm. And in fact the U.S. military plans to use robotic vehicles as teammates or “robotic wingmen” for troops on the frontlines. This approach is often described as “manned-unmanned teaming” and is envisioned for a number of contexts, including frontline air and ground combat. In some cases, these robotic systems are intended to carry additional weapons or gear. In other cases, they may be operating as forward scouts, reducing the danger to humans on the ground.
More generally, the idea that drone pilots have no skin in the game taps into a complex relationship war fighters have with danger. Risk is a relative concept, and it is common in military circles to denigrate others who are exposed to less harm. To the fighter pilot, a drone pilot sitting in a trailer stateside may not be a true warrior. (Indeed, the disparity in medals given to drone pilots speaks to their treatment as second-class citizens in the air force.) To the grunt in the mud dodging enemy bullets, the pilot zipping overhead is hardly exposed to the same risk. Those who venture outside the wire in modern wars look down on “fobbits,” who rarely leave the confines of forward operating bases (FOBs). Combat troops look down on support troops, even those who go on patrol. Light infantry troops sneer at mechanized infantry, who hitch a ride to battle. Airborne troops scoff at “legs,” who don’t jump into battle.
Troops actually embrace robotic systems if they help reduce their own exposure to harm.
These perspectives are timeless. The crossbow was decried as an immoral weapon when it was created, in part because it allowed less skilled and heroic archers to kill knights from a distance without exposing themselves to harm. In time, warfare adapted to killing from a distance. Today’s commandos fearlessly charge into battle, but ancient warriors who fought hand-to-hand with swords and axes would likely see shooting someone from a distance as cowardly. Does that mean we should give up airplanes and rifles and fight, bravely, with knives?
The reality is that ideas about bravery shift over time. The long arc of military technologies has been toward killing at greater and greater distances, from the sling and the stone to cannons, missiles, and drones. The advantages of adopting a new technology first—such as those the United States enjoys today with drones—do not last long. New weapons proliferate, resulting in a new battlefield equilibrium. Terrorist groups are already using their own drones in Iraq and Syria. Fighting remotely may be seen as cowardly today, but these attitudes are likely to change, either through peacetime innovation or in the crucible of combat. On the battlefield, combat effectiveness will trump bravery every time.
Early versions of technologies, moreover, often have significant shortcomings that make their military utility questionable. First-generation matchlock rifles, tanks, and aircraft had major limitations but improved over time. One of the most significant drawbacks of today’s drones is their limited automation. U.S. Defense Department road maps for robotic systems have again and again highlighted the importance of greater autonomy. As these systems become more automated and shift away from today’s remote control paradigm to a more blended human-machine teaming approach, many of today’s problems, such as fragile communications, are likely to improve. Other shortcomings of robotic systems may endure, but their advantages may still outweigh their limitations. Looking at first-generation tanks, one might have easily concluded that horses remained a better option for cavalry. Tanks require fuel and oil and they break down easily, while horses are robust and can eat off the land. This remains true today! Yet the advantages of tanks overwhelm these shortcomings.
THE ROBOTICS REVOLUTION
Over 90 nations and nonstate groups already have drones, and militaries around the world are racing to develop ever more sophisticated robotic systems. The robotics revolution is happening regardless of whether or not the U.S. military leads the way.
Cultural resistance to robotic systems in the U.S. military is real. As Schneider and Macdonald highlight, it is not always confined to those whose jobs are at stake. Some of this resistance may be due to real limitations of drones today, and engineers should work with war fighters to understand their concerns and improve next-generation systems. Other objections may lie closer to the backward resistance seen in past generations to crossbows, tanks, and other innovations. The U.S. military will face a future with robotic weapons on the battlefield. The only real choice is whether they are the United States’ robots or the enemy’s.
The robotics revolution is happening regardless of whether or not the U.S. military leads the way.
SCHNEIDER AND MACDONALD REPLY
We appreciate engagement on this issue and Paul Scharre’s thoughtful response. Scharre’s comments represent an important move forward in the debate about the future role of autonomous systems on the battlefield, and one that we welcome.
Our disagreement seems to be that Scharre is more confident than we are that technological advances will solve the trust issues that we identify. It is important to note that we are not saying such developments will neverhelp—there is inevitably a tipping point where confidence turns into trust—but simply that we are less confident that this will happen in the short term or that they will solve all the issues that troops identified during the course of our research. We all agree on the need to establish where this tipping point is and how and when it might be reached.
As for our policy recommendations, we are not recommending that the U.S. military give up on unmanned technologies altogether—only that it cannot use unmanned technologies to revolutionize warfare if it focuses only on technology and fails to address the human side of the equation. For instance, should the United States try to improve drones’ combat effectiveness by building better sensors, deploying drone operators to remote locations, allowing troops to control the system from the battlefield, or better integrating drones in large-scale ground exercises? The answer is probably some combination of all of the above, but right now the preponderance of effort is on the engineering solution at the expense of the experiential.
As Scharre notes, the U.S. military will undoubtedly face a future with robotic weapons on the battlefield. The key question is what kinds of roles these technologies will play. The third offset, a U.S. Defense Department acquisition strategy put forward under former Secretary of Defense Ashton Carter to “offset” rising military capabilities in countries such as China and Russia, identifies autonomy as the key technology for winning future wars. This offset strategy frames machine decision-making as a capability that will necessarily create revolutionary advantages for states that utilize autonomy on the future battlefield.
However, advocates of the third offset have had trouble articulating how autonomy translates to the battlefield advantages they predict. For the first offset, the United States focused on developing larger-yield nuclear warheads and a massive retaliation doctrine in order to offset Soviet numerical superiority. For the second offset, the United States used the advent of the microprocessor to leverage precision technologies and network-centric operations in order to offset Soviet advantages in forward presence and mass. But which tenets of warfare does autonomy enhance? Which types of campaigns do they benefit?
Without an understanding of how these unmanned and increasingly autonomous systems affect principles of war, the debate about unmanned systems often reverts back to arguments that cannot be disproven. Opponents of these systems can argue that they do not provide new capabilities not already provided by manned systems. On the other side, unmanned proponents can say that the systems will provide revolutionary battlefield capabilities but have trouble explaining how. We need to better understand what capabilities unmanned systems can create in order to achieve victory; being autonomous in itself is not a capability.
Finally, our current survey data are, like all survey data, limited by the technology available at the time. It would be fascinating to do follow-up work to see whether, for example, the DARPA Persistent Close Air Support program would reduce objections, as Scharre suggests it might. It would also be useful to try to understand more about how different groups perceive unmanned systems. Not everyone who opposes unmanned is a pilot protecting his or her turf and not everyone who supports unmanned is sitting at DARPA or behind a drone terminal. What explains the motivations for developing or resisting unmanned systems? Would we see less resistance to underwater unmanned vehicles? In space? Autonomous cyber-operations? These are all questions for future research and policy.