Headlines would lead one to conclude that the age of killer robots has arrived. In Ukraine, autonomous drone swarmssupposedly hunt for enemies and independently decide what to attack. Because Ukraine and Russia are employing many different types of drones, which are omnipresent over the frontlines, some observers have concluded that drones have fundamentally changed the character of war and are the key factor that will decide who prevails in the conflict. In my new Center for a New American Security Report, Evolution Not Revolution: Drone Warfare in Russia’s 2022 Invasion of Ukraine, I conclude the opposite. Drones have not fundamentally altered the character of war and will not determine who wins or loses this war. Nevertheless, drones are changing how Ukrainian and Russian troops fight at a tactical level and will impact all future battlefields.
Drones have indeed transformed the battlefield in Ukraine by providing accessible and affordable capabilities at a scale that did not previously exist. They are making it difficult to concentrate forces, achieve surprise, and conduct offensive operations. While drones are not more survivable than crewed aircraft, they enable greater risk acceptance. Moreover, drones do not have to be survivable if they are cheap and plentiful as they can achieve resiliency by reconstitution. Nevertheless, the overall impact of drones has been more evolutionary than revolutionary. Drones connected to ground-based fires units have made common artillery shells precision weapons. First-person-view kamikaze drones can accurately hit mobile targets, making the frontlines even more lethal. Nevertheless, even large numbers of small drones cannot match the potency or volume of artillery fire and thus cannot substitute for howitzers. Also, while drones provide affordable airpower, they have not replaced traditional air forces, nor have they been able to obtain air superiority. With this in mind, Washington should continue helping Ukraine improve its drone fleet, while being realistic about the impact this will have.
These Are Not the Drones You Are Looking For
Distinguishing fact from fiction during a war is often difficult, particularly when both sides are actively engaging in information warfare to shape others’ perceptions. The first drone that comes to mind for many when they think about the war in Ukraine is much-hyped Turkish built Bayraktar TB2, which was active in the first days and weeks of the war and helped to repel Russia’s initial attack. Yet overall, medium-altitude long-endurance drones, like the TB2 or the Russian Orion, have not played a large role in this war. Turkish propaganda portrayed the TB2 as a fantastic weapon that could evade even advanced air defenses and was affordable. In reality, however, once Russia relaxed the rules of engagement for its surface-to-air missiles most of the TB2s were quickly shot down and they essentially disappeared from the battlefield.
While large reusable military drones are rarely employed in this war, Ukrainian and Russian ground forces are making extensive use of smaller military and commercial or do-it-yourself drones. Military intelligence surveillance and reconnaissance drones, such as the Ukrainian Furia or Flyeye or the Russian Orlan-10 or ZALA 421 variant, are more capable and costly than most of their commercial counterparts and thus are usually commanded by larger units, like battalions. As seen Figures 1 and 2, these military drones fly farther, can stay aloft longer, and using their sophisticated sensors offer persistent surveillance.
In contrast, small commercial quadcopters, which only cost a few thousand dollars, are ubiquitous, but have shorter ranges and limited battery life as seen in Figure 3. Commercial drones are providing intelligence to frontline soldiersabout the battlespace that previously was only available to higher headquarters. Thus, no Ukrainian or Russian unit would try to maneuver or launch an assault without at least one small commercial drone to scout and provide real-time information on the disposition of nearby enemy and friendly forces. Because Russian and Ukrainian ground units at every echelon have multiple commercial drones in their inventory, the skies above the front lines are blackened with small quadcopters.
The net effect of many small military and commercial drones on the war is to make the defense dominant and the frontlines extremely lethal. Consequently, it is difficult for Ukrainian and Russian forces to move or mass, and thus to achieve tactical surprise or to go on the offensive. But, as Franz Stefan-Gady has explained, offensive operations are still feasible if one side takes advantage of environmental factors that inhibit drone operations. The frontlines in Ukraine are not impenetrable or completely transparent, but drones are forcing troops to disperse and conceal themselves, which has overall made it harder to maneuver and attack.
The accessibility and affordability of drones is creating new capabilities at a scale that previously did not exist, but it is not drones alone that are transforming the battlefield. Instead as Mick Ryan and Clint Hinote have argued it is the combination of large numbers of weapons systems linked together that is truly disruptive. In Ukraine, the dominant pairing has been unarmed drones and artillery, which have accelerated targeting timelines and enabled responsive and precise ground-based fires. As Jeffrey Edmonds and Samuel Bendett have highlighted, drones are the key links in what Russia calls its reconnaissance-strike complex, or the network of forces that conduct targeting for artillery units. As artillery firepower has been the dominant weapon of this war, drones have played a critical enabling role as spotters that acquire targets and help to adjust fires by communicating targeting information using virtual battle networks, such as Kropyva and Strelets, to fires units. Oftentimes, it is multiple drones operating in stacks in the same airspace that perform different roles, creating a highly distributed and resilient kill chain. One high-flying drone initially identifies the potential target and then sends in a cheaper smaller drone to verify that it is an enemy formation worth engaging. A third drone acquires the targeting quality data and passes it to the howitzers that then fire. The same drone responsible for targeting or another small drone may assess if the target was destroyed and if not help to adjust the aimpoint. In this way, drone spotters enable imprecise indirect fire weapons to have precision effects.
The Rise of the Kamikaze and Do-It-Yourself First-Person-View Drones
Additionally, in the last year kamikaze drones or loitering munitions have provided a precision-strike capability, especially when paired with unarmed drones and artillery. Kamikaze drones are one-way systems that crash into their targets and are more akin to a munition than an aircraft that is supposed to fly many sorties. Military-grade kamikaze drones, such as the Ukrainian Switchblade 600 or Russian Lancet-3, have become the preferred weapon for counterbattery fire. When an unarmed drone, such as a Puma or Orlan-10, finds an enemy artillery unit, a kamikaze drone is then launched to destroy the gun or radar. Thus, military kamikaze drones and surveillance drones are a key factor in determining who wins an artillery duel.
In the last six months, do-it-yourself first-person-view kamikaze drones flown by operators wearing goggles have proliferated. Ukraine was the first to assemble these cheap weapons from commercial racing drones, but Russia was a fast follower and its production of first-person-view drones now outstrips that of Ukraine. At approximately $400 each, do-it-yourself kamikaze drones cost a fraction of other weapons and allow soldiers to attack moving targets beyond their line of sight. First-person-view drones offer cheap precision anti-personnel and anti-tank capabilities, but they remain tactical weapons with limited range, even with the help of drone communications relays. Typically, these racing drones operate with unarmed surveillance drones that have more endurance and thus can find the targets, conduct battle damage assessment, and direct the fast-flying kamikaze towards the target before its battery dies.
First-person-view kamikazes were an unexpected development fueled by commercial technologies that have disrupted the battlefield because they are so cheap and thus abundant. Yet do-it-yourself drones by themselves are not substitutes for existing weapons systems. These kamikaze drones have significant limitations and are most effective when operating with unarmed drones and artillery. While first-person-view drones have a longer range than most anti-tank weapons, they carry smaller payloads than the American Javelin. Typically, first-person-view drones disable or achieve a mobility kill against tanks vice destroying the vehicle with one strike. Once a tank has been immobilized, it then is destroyed by artillery fire or follow-on kamikaze attacks that focus on weak points, such as the gas tank or the rear of the turret, which if pierced are likely to cause a larger explosion that can destroy the vehicle. These tactics employing first-person-view drones, surveillance drones, and artillery have proven very effective and, as Michael Kofman and Rob Lee have noted, forced forces near the frontlines to eschew the use of vehicles in favor of distributed dismounted maneuver and assaults. Do-it-yourself kamikaze drones, therefore, are an effective complement to existing weapons that can help to conserve scarce artillery ammunition and supplement offensive firepower on the frontlines.
Importantly, even large numbers of small do-it-yourself kamikazes cannot match the potency of artillery fire. Artillery shells pack a bigger explosive punch and can be rapidly fired in large salvos. Howitzers also can provide sustained fires at a slower but continuous rate. The basic U.S. 155mm shell, for example, contains almost 24 pounds of explosive compared to the three pounds carried by a first-person-view drone. Large kamikaze attacks consisting of many first-person-view kamikazes currently are limited in size due to electronic interference and because each drone must be individually controlled and actions between drones manually coordinated. Thus, first-person-view drones cannot currently produce the volume of fires of artillery salvos.
Moreover, racing drone attacks are limited by the need for very skilled pilots. Because first-person-view drones are not fire-and-forget weapons, they require an operator with the acuity to quickly identify weak points on a target and the physical dexterity to guide the kamikaze into the exposed point. Unlike basic commercial quadcopters, not everyone can acquire these skills. Moreover, even those who have the core capabilities need weeks if not months of training to become truly proficient.
The fact that first-person-view drones remain tethered to their human pilot is also a liability on the battlefield. While racing drones are difficult to physically intercept, they can be stopped by jamming the control link. Ukrainian developers have been working on creating software that enables the first-person-view drone to lock onto its target and complete its kamikaze mission even if the control link is cut. This would be a notable breakthrough that could negate one of the most effective defenses against first-person-view drones and potentially reduce the amount of skill that is needed to pilot them. In other words, if this breakthrough were to occur, it might overcome training bottlenecks and enable do-it-yourself kamikaze drones to scale more effectively. Yet the use of artificial intelligence and autonomy in the Ukraine war is another topic that is often misunderstood.
Armies of Drones Tethered to Humans
Despite many headlines about autonomous Russian and Ukrainian drones, discussions about artificial intelligence and autonomy generally lack precision and as a result create false impressions about the level and type of autonomy that exists on the battlefield. The vast majority of drones in the war in Ukraine are remotely piloted and humans not machines remain the interface that manually coordinates the actions of multiple drones. Thus, there are no true drone swarms or cooperative autonomy. Drones usually operate together in stacks with pilots communicating via text chat in a virtual battle network or via cell phones. As Jack Watling has highlighted, uncrewed weapons are still people-centric. Similarly, Stefan-Gady has observed that Ukrainian assault squads have nearly the same number of drone operators as infantrymen. Each drone usually has two human operators, one who is focused on navigation and communications and the other to steer the aircraft. Even though most drones remain piloted by humans, limited forms of autonomy abound. For instance, many military and commercial drones autonomously perform mundane but important tasks, such as automatic take-off and landing, and have collision-avoidance software built in.
These types of autonomous capabilities are critical but are a far cry from the fully autonomous drones that independently make decisions about who, what, and when to kill. According to Paul Scharre, the author of Army of None: Autonomous Weapons and the Future of War and Four Battlegrounds: Power in the Age of Artificial Intelligence, fully autonomous weapons “are defined by the ability to complete the engagement cycle—searching for, deciding to engage, and engaging targets—on their own.” Ascertaining whether a weapon autonomously makes decisions about life and death is particularly challenging because it is unobservable except to the drone operator. A drone dropping a bomb or crashing into the target looks the same whether it was remotely controlled with an operator selecting the target and deciding when to engage it or if the drone had the capability to autonomously identify targets and to strike them independently. Oftentimes, as was the case with the Russian Lancet-3 kamikaze drone, one discovers that a weapon’s manufacturer has exaggerated the capabilities of the system, which does not have the level of autonomous capability promised or perform as expected.
Other times, the degree of autonomy is circumscribed, and the drone has a form of narrowly bounded autonomy that is common in other types of weapons. The type of autonomy that Ukraine is trying to develop for its first-person-view kamikaze drones is an example of this. For its do-it-yourself kamikaze drones, Ukrainians are building software to autonomously perform object recognition and navigation. These capabilities already exist in military kamikaze drones like the Switchblade 300 or other homing weapons like the Javelin or the AIM-120 Advanced Medium Range Air-to-Air Missile. The human selects the target and then the weapon automatically guides itself to the identified target. Fire-and-forget terminal guidance is tightly bound in time and space and should not be feared. If the Ukrainians can develop this type of software cheaply using commercially available systems, it would be another major step in the democratization of precision strike, but not a fully autonomous killer robot.
There are signs that Ukraine and Russia are working on developing fully autonomous drones, but these have not been widely fielded — likely because they are not yet effective. The Ukrainian Saker Scout drone supposedly has object recognition capability and aspires to be able to act fully autonomously. David Hambling reports that the Saker Scout can identify 64 different targets and that it has been used in a fully autonomous attack. Yet today fully autonomous drones seem to be an exception rather than the norm in Ukraine, which suggests that the fully autonomous systems are very expensive or do not work as promised. In all likelihood, it is the latter. Object recognition is easier in uncluttered environments, like the air or sea, but much more difficult on the ground when an enemy is trying to hide. On the frontlines in Ukraine, where there is little standardization in terms of equipment, it would be particularly difficult to develop a fully autonomous system that could recognize all of the different variations of military equipment that was also trusted by its users. As artificial intelligence matures, fully autonomous drones may be fielded in the future. But it would likely be a more incremental process that keeps a human in or on the loop and allows soldiers to refine and develop trust in the system. For instance, an unarmed drone tasked with target acquisition may use artificial intelligence to identify targets that can then be verified by a human and engaged by an armed drone.
In sum, many drones in Ukraine have different types of limited autonomous capabilities and Ukraine may be close to creating semi-autonomous first-person-view drones, but there are few if any fully autonomous drones in this war. Given the rate at which artificial intelligence is developing and the pace of innovation on the battlefield, this could change in the next year or two, especially since fully autonomous drones could provide important tactical advantages.
Evolution in Military Affairs
Individually all of these drone innovations are notable advancements, but even cumulatively they do not add up to a revolution. According to Andrew Krepinevich, revolutions in military affairs must “fundamentally alte[r] the character and conduct of a conflict” by “producing a dramatic increase—often an order of magnitude or greater—in the combat potential and military effectiveness of armed forces.” Revolutions in military affairs are so disruptive that they render old weapons, ways of fighting, and organizational constructs obsolete. Thus, revolutions require more than widespread adoption of new technologies. Additionally, militaries must develop new operational concepts, integrate new capabilities into broader military systems, and adapt their organizational culture and structure. Not surprisingly, this wholesale sort of change takes time. President Volodymyr Zelenskyy recently announced the creation of a separate Ukrainian drone service, which could be the harbinger of a revolution. Yet this is far from certain. Because drones in the Ukraine war have been most effective when they are combined with other capabilities, creating a separate service could create more barriers to effectively integrating drones with other weapons. In many ways this is an academic debate about what constitutes a revolution in military affairs, which is irrelevant to the larger point that the accessibility and affordability of military drones and cheap commercial drones are profoundly changing warfare at a tactical level in Ukraine. But it is important not to exaggerate the impact of any one technology or weapons system because there is no one technological silver bullet that will win this conflict for Ukraine.
Because most of the drones in Ukraine are commercially derived systems, the technology has quickly diffused to the enemy and has not provided an enduring advantage to either side. Throughout the war, there have been rapid cycles of adaptation as both sides have learned from each other, adopting tactics and technologies that have been used successfully and developing counters to improve their defenses.
This pattern is likely to continue as the war drags on, but there are steps that can be taken to bolster Ukraine’s drone capabilities and harness its people’s technical ingenuity. First, the United States and Europe should help Ukraine to develop software that would enable semi-autonomous first-person-view drones. Because Russia has been less adept at incorporating commercially derived software into its weapons and operations, do-it-yourself kamikaze drones with autonomous terminal guidance could provide Ukrainian forces with a more durable advantage. Second, the United States and Europe need to help Ukraine to evolve from an artisanal start-up model of production to industrial-scale drone manufacturing, especially for larger kamikaze drones and first-person-view drones. Russia has surpassed Ukraine in the drone fight because of its ability to scale production of small military drones. On a one-for-one basis, Russia’s drones are often inferior to Ukrainian drones, but Russian forces simply have deeper inventories of these critical weapons. Ukraine can and should expand its capacity to produce drones, which will necessitate picking winners and losers and scaling production of the best systems.
Over time, as drones become more autonomous and are more broadly connected with other weapons, they may fundamentally reshape military doctrine and organizations and truly revolutionize warfare. But thus far in Ukraine, drone warfare has been an evolution, not a revolution. It is clear that drones alone will not determine who prevails in this conflict, but they will certainly play a prominent role in the ongoing war in Ukraine and in other battlefields in the future.
Stacie Pettyjohn is a senior fellow and director of the Defense Program at the Center for a New American Security.
Image: Ukrainian Ministry of Defense
Comments are closed.