Next-Generation Drone Warfare: Interview with Paul Scharre
Editor’s Note: The below is an RCD Interview with Paul Scharre, a Fellow and Director of the 20YY Warfare Initiative at the Center for a New American Security (CNAS). From 2008-2013, Mr. Scharre worked in the Office of the Secretary of Defense (OSD) where he played a leading role in establishing policies on unmanned and autonomous systems and emerging weapons technologies.
Prior to joining OSD, Mr. Scharre served as a special operations reconnaissance team leader in the Army’s 3rd Ranger Battalion and completed multiple tours to Iraq and Afghanistan. Mr. Scharre has published articles in Proceedings, Armed Forces Journal, Joint Force Quarterly, Military Review, and in academic technical journals.
Assistant Editor John Waters: In CNAS’ recent working paper, “An Introduction to Autonomy in Weapon Systems,” an autonomous system is defined as a machine that performs some task or function on its own. Based on the complexities of robotic judgment and decision-making, do you foresee the greatest potential for autonomous drones in force protection, surveillance and reconnaissance, precision targeting, logistics/troop transport, command and control (C2) or combat roles? Why?
Paul Scharre: I think autonomous systems will be useful for all of these tasks. Autonomy shouldn’t be thought of as a niche feature. It will be fundamental to future military systems, like communications or propulsion. The industrial revolution brought about powerful machines on land, in the air, and on and under the sea. The internal combustion engine wasn’t limited to only some military roles. We’re in the middle of an analogous shift today. The information revolution is making machines – both hardware and software – smarter. That will be useful in nearly every aspect of military operations.
Waters: Operational and tactical military commanders have become reliant upon the high-definition, full-motion video (FMV) provided by theater-level intelligence, surveillance and reconnaissance (ISR) platforms and made popular by Predators and Reapers. Will the data link of these next-generation automated systems be able to provide this particular feature in a contested environment against a first-world adversary?
Scharre: Absolutely not. There will be no way in a contested environment to stream high-definition, full motion video. The good news is that you don’t have to. The command and control links to control uninhabited vehicles are an order of magnitude less than streaming full motion video and are much more manageable. Remote human operators will still need to get data from uninhabited vehicles in order to make decisions, such as which targets to engage. But onboard data processing can help to cue those targets to human operators, so the human operators don’t need to be streaming continuous video. Rather, autonomous data processing tools onboard the vehicle can say, “this might be a target, take a look,” and send a snapshot to a person, who then makes the call. This is much more manageable from a bandwidth perspective.
For a specific example, it might be worth checking out DARPA’s CODE project, which is examining this exact issue – collaborate teaming of autonomous systems in contested environments.
Waters: How troubling is the current divide between vision and funding for these next-generation systems?
Scharre: There are two hurdles. One of them is funding. Right now, about 1 in every 20 U.S. defense procurement and R&D dollars is spent on uninhabited, or unmanned, systems. That isn’t enough and is not reflective of their potential in future military operations. But perhaps the bigger hurdle is conceptual. All of the history of innovations in warfare suggests that it isn’t the widget itself that leads to disruptive change on the battlefield, but rather uncovering new ways of fighting. It’s the doctrine, organization, and concepts of operation that really lead to a significant change in warfare. Achieving that requires more than just a new gadget – it requires an iterative process of experimentation and concept development, alongside further refining the technology. It also requires the willingness to change our paradigm for operations … and that’s the hard part.
Waters: Whether a Joint terminal attack controller (JTAC) providing target guidance to an overhead attack aircraft, navy watch officer navigating an amphibious ship during a replenishment at sea (RAS) or an army platoon leader issuing combat orders to an Afghan commando unit, operational command and control (C2) is among the greatest challenges for any military force. How significant will the challenge be to C2 autonomous drone swarms on an already crowded – and often coalition or multi-national – battlefield?
Scharre: Swarming will force us to rethink our command-and-control paradigms. Right now we have a paradigm where we are remotely controlling uninhabited, or unmanned, systems. In fact, the Air Force goes to great lengths to call their uninhabited aircraft “remotely piloted.” But as future systems incorporate greater autonomy, we shouldn’t have people “piloting” or “driving” aircraft or vehicles in the traditional sense of directly controlling their movements. That should be automated. Rather, people should be commanding them at the mission level, directing the vehicle where to go and what task to perform. This is already the case underwater where uninhabited vehicles require a high degree of autonomy because the communications environment is so challenging. Swarming takes this even another step further. With swarms, the person isn’t tasking any one vehicle, but rather tasking the swarm as a whole, which then coordinates among the different swarm elements to accomplish a task. This requires thinking about command and control very differently, and it isn’t yet clear what the best way is to command and control a swarm. Figuring that out will require experimentation, and the best model may depend on the environment and the mission.
There may not be any one universal answer. Examples from nature can help, since there are many examples of swarming behavior, from ants and bees to wolf packs, but there are fundamental differences between natural swarms and robot swarms. For one, robot swarms are designed, not evolved. They have a specific purpose. But more importantly, for robot swarms there will always be a human controller in charge at some level. Even if the human operator lacks communications with the swarm once it is launched, he or she is launching the swarm to perform a specific task for a specific purpose. There isn’t a similar analogy in nature. So it’s a very new area, and there is interesting research ongoing in military and academic labs.
Waters: Combined arms warfare integrates various elements of the combat arms – airpower, artillery, armor and infantry, to name a few – to achieve complementary effects against an enemy force. In most cases, combat arms assets support infantry troops on-the-ground as they assault an objective or capture key terrain. Do you foresee a reversal of this relationship: A future battlefield where a small contingent of multi-skilled infantry or special operations troops supports autonomous drone warfare?
Scharre: I’m not sure exactly how it will look, but in general I don’t foresee a future battlefield devoid of humans, with just robots fighting robots. I think there will still be humans forward in the battle space, but probably more in a command-and-control context. The types of things humans are doing will change. And it may be robots on the front lines, but that doesn’t mean that humans on the battlefield will be safe. Bullets and missiles gave humans greater standoff from the enemy, but they haven’t necessarily made war safer, because the enemy has them too. Remember, the enemy will also have autonomous swarms coming after you.
But to your point about combined arms – I do think our vision of combined arms operations will shift over time. Just like the advent of the tank, aircraft, and radio brought about the blitzkrieg, the way we think about maneuver operations will need to change. Perhaps ground maneuver warfare will be replaced with the air-mobile swarm … I don’t think we yet know. The key thing is a process of experimentation to figure out the best ways of employing this technology and being willing to change our concepts of operation to make the most of it. I don’t yet see that happening. For example, the Army is starting development of a light airborne tank.
But does that even make sense in an era of uninhabited ground vehicles? Uninhabited ground vehicles allow us to break the “iron triangle” of lethality, protection, and mobility. We can build heavily protected vehicles that have people in them and then have lighter, faster uninhabited vehicles with less armor for scouting, deception, and the front line of an advance. Maybe that’s where we should be going. The natural instinct of DoD is to build a next-generation weapon system that looks like what we already have, but better. A better tank, better ship or better fighter. But what if the future of warfare is something different? You don’t want to be caught working on better stirrups for horse cavalry when the enemy is inventing the tank.
Waters: Despite the apparent ‘end’ to U.S.-involvement in the wars in the Middle East and Southwest Asia, the United States continues to wage a semi-secret war in the Arabian Peninsula and Horn of Africa using unmanned aerial combat vehicles. Is the current rate of development in drone technology out-pacing our ability to regulate UAV employment with regard to the law of armed conflict and rules of engagement?
Scharre: The U.S. government has acted as though it has a monopoly on drones. It doesn’t. North Korea has flown small drones into South Korea. Hezbollah and Hamas have flown them into Israeli airspace. Allegedly both Russian and Ukrainian forces are flying drones in the Ukraine. China is developing a stealth drone.
This change in the U.S. drone export policy (outlined on Feb. 12 in The Washington Post) has been long overdue. The United States has refrained, to-date, from exporting unarmed drones to key partners like Jordan and armed drones to even close allies like Italy and France. That doesn't make sense, given that we already supply some of these countries very sophisticated military equipment. It also is in U.S. interests to strengthen partner capabilities in this area to help offset the U.S. shortfall in airborne surveillance against groups like ISIS.
The idea of coupling transfers with principles for proper use is key, as it gives the United States the opportunity to shape expectations for how drones are used internationally. Drone technology is proliferating abroad no matter what the United States does. What countries will do with them is the question. Conditional transfers to key partners and allies who will use them in ways that are legitimate and lawful are important tools in shaping expectations for appropriate use.
Waters: What is your take on the current threat of drone proliferation to non-state actors, trans-national terrorist groups and violent extremist organizations? How must the United States adapt to the challenges posed by drone proliferation?
Scharre: Cheap drones are widely accessible online, and while they don’t pack as much punch as a Reaper, they could definitely do some damage. I think as recently as a couple years ago many defense analysts dismissed this threat, seeing cheap off-the-shelf drones as not significant in the traditional military sense. And in the traditional sense, they aren’t. But they could still do quite some damage. So I think that perspective is changing and people are starting to wake up to this threat. In particular, the situation where for the past sixty years U.S. ground forces have not had to face threats from the air is likely ending. The type of challenge we’ve seen with IEDs in the last wars we could likely see in the air coming at us in the next war.
A future of ‘unmanned’ drones fighting our battles in the skies, under the sea and on land is largely misleading. According to Scharre, the words we use to describe next-generation drones affect our ability to comprehend their future place on the battlefield, and understand their functionality in proper context.
In short, humans will always be in control at some level. In both the physical and cognitive senses, future warfare will likely feature ‘human-machine teams’ performing several mission-specific functions under the direction of human commanders. Just as an infantry rifle squad is task-organized to ‘locate, close with and destroy by fire and maneuver,’ so too will a swarm of autonomous drones execute tasking to conduct a movement to contact, to attack an enemy C2 node or to reconnoiter a route for overland logistics transport. And if the communication or data link between robot and commander should fail during the mission, the autonomous systems will have programmed instructions to follow contingency tasking, such as land at a pre-designated GPS coordinate.
The word ‘autonomous’ should not be interpreted to mean untethered from human control. Rather, we should begin envisioning a battle space where automated systems perform – not just enable – critical war fighting functions at the behest of human leadership.