Obstructive warfare: Applications and risks for AI in future military operations

Fox, Amos C. Working Paper Obstructive warfare: Applications and risks for AI in future military operations CIGI Papers, No. 307 Provided in Cooperation with: Centre for International Governance Innovation (CIGI), Waterloo, Ontario Suggested Citation: Fox, Amos C. (2024) : Obstructive warfare: Applications and risks for AI in future military operations, CIGI Papers, No. 307, Centre for International Governance Innovation (CIGI), Waterloo (Ontario) This Version is available at: https://hdl.handle.net/10419/306703 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by/4.0/ CIGI Paper No. 307 — October 2024 Obstructive Warfare Applications and Risks for AI in Future Military Operations Amos C. Fox CIGI Paper No. 307 — October 2024 Obstructive Warfare Applications and Risks for AI in Future Military Operations Amos C. Fox About CIGI The Centre for International Governance Innovation (CIGI) is an independent, non-partisan think tank whose peer-reviewed research and trusted analysis influence policy makers to innovate. Our global network of multidisciplinary researchers and strategic partnerships provide policy solutions for the digital era with one goal: to improve people’s lives everywhere. Headquartered in Waterloo, Canada, CIGI has received support from the Government of Canada, the Government of Ontario and founder Jim Balsillie. À propos du CIGI Le Centre pour l’innovation dans la gouvernance internationale (CIGI) est un groupe de réflexion indépendant et non partisan dont les recherches évaluées par des pairs et les analyses fiables incitent les décideurs à innover. Grâce à son réseau mondial de chercheurs pluridisciplinaires et de partenariats stratégiques, le CIGI offre des solutions politiques adaptées à l’ère numérique dans le seul but d’améliorer la vie des gens du monde entier. Le CIGI, dont le siège se trouve à Waterloo, au Canada, bénéficie du soutien du gouvernement du Canada, du gouvernement de l’Ontario et de son fondateur, Jim Balsillie. Copyright © 2024 by the Centre for International Governance Innovation The opinions expressed in this publication are those of the author and do not necessarily reflect the views of the Centre for International Governance Innovation or its Board of Directors. For publications enquiries, please contact [email protected]. The text of this work is licensed under CC BY 4.0. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. For reuse or distribution, please include this copyright notice. This work may contain content (including but not limited to graphics, charts and photographs) used or reproduced under licence or with permission from third parties. Permission to reproduce this content must be obtained from third parties directly. Centre for International Governance Innovation and CIGI are registered trademarks. 67 Erb Street West Waterloo, ON, Canada N2L 6C2 www.cigionline.org Credits Managing Director and General Counsel Aaron Shull Director, Program Management Dianna English Program Manager and Research Associate Kailee Hilt Publications Editor Christine Robertson Publications Editor Susan Bubak Graphic Designer Sami Chouhdary Table of Contents vi About the Author 1 Executive Summary 1 Introduction 3 Stand-Off Warfare and the Limitations of Technology 6 Reflections on Stand-Off Warfare in the Russia-Ukraine War 8 The Logic of Land War 10 The Potential Impacts of AI on Military Operations 15 Policy and Interoperability 16 Conclusion: Policy Recommendations 18 Works Cited vi CIGI Paper No. 307 — October 2024 • Amos C. Fox About the Author Amos C. Fox is a retired US Army officer with multiple combat deployments to Iraq. Amos has a Ph.D. in international relations from the University of Reading and he is a lecturer in political science at the University of Houston and a contributing editor at War on the Rocks. He also hosts the Revolution in Military Affairs podcast and War on the Rocks’ Soldier Pulse and WarCast podcasts. Amos is the author of Conflict Realism: Understanding the Causal Logic of Modern War and Warfare (2024, Howgate Publishing). 1Obstructive Warfare: Applications and Risks for AI in Future Military Operations Executive Summary Artificial intelligence (AI) provides seemingly limitless potential for applications in contemporary and future armed conflict. Optimistic futurists claim that AI might transform war into a sanitized situation in which civilian casualties and collateral damage can be almost zeroed out from the battlefield. These futurists also posit that AI can amplify the speed at which military leaders and policy makers can gain situational understanding and make prudent decisions at the pace of information. In doing so, futurist theory predicts that militaries, using dual-use cyber and spaced-based remote and deepsensing capabilities, will make “kill chains” that provide an asymmetric advantage over potential adversaries; thus, whoever masters the kill chain concept will dominate future wars. More conservative estimates suggest that AI can open new horizons in war and warfare, but not to the extent advocated by many optimistic futurists. This paper sides with this more conservative perspective by asserting that AI’s transformative impact on the future of war will fall short of the predictions of futurists. The futurists’ vision of future war, and their advocacy of stand-off warfare, is incompatible with the fundamental and inherent challenges of land war. Coupled with the increasing number of network, global and geospatial sensors, AI will provide policy makers and military leaders with massive amounts of information. But this excess of information, while seemingly a valuable asset, is equally a vulnerability. Strategic actors understand that increased battlefield and global sensing will also work to deceive sensing and inject incorrect data into an adversary’s information network. Moreover, AI’s contributions to battlefield and global sensing will make movements on future battlefields increasingly deadly for armed forces; as a result, military forces will likely embrace positional warfare and revert to operating in and around urban areas, where both state and non-state forces will cause more civilian casualties and collateral damage in future wars. Therefore, policy makers, military leaders and scholars should anticipate AI increasingly contributing to data pathway warfare, in which combatants use information in innovative ways to overcome remote, deep and battlefield sensing capabilities. AI’s ability to make the battlefield more transparent for policy makers and military leaders might result in both military and non-state military forces adopting positional warfare to offset the advantages that AI-enabled sensing provides to combatants. Policy makers, military leaders and scholars should also anticipate an increase in urban warfare as combatants — both state and nonstate — seek to offset the potential speed that AI might bring to sensor-to-shooter kill chains. When viewed collectively, these transformative aspects of AI will potentially result in longer conflicts; attritional wars, with increased civilian casualties and collateral damage; and munitions shortages, if industrial bases are not retooled to keep pace with the potential speed of future kill chains. Introduction This paper introduces a new theory of warfare— obstructive warfare — to sidestep the wide-ranging sensationalism associated with today’s new and emerging technology and instead provide an alternative assessment, based in causal logic, for how AI can be used in military operations. Obstructive warfare is anchored in the belief that all land wars carry with them a set of nearly unavoidable challenges, which are outlined later in this paper. Military forces cannot overcome these challenges solely with “attacks from above,” nor can battlefield transparency prevent these challenges from materializing. The challenges of land war often cause states to fight wars positionally, or through the purposeful use of movement in combination with location(s) to dislocate an adversary’s strength, accentuate one’s own power and generate favourable situational warfighting asymmetries to defeat or destroy the adversary (Fox 2017, 18). Considering positional warfare’s proclivity for forceoriented military operations that use movement, location and the application of power (i.e., military firepower), it is easy to understand how this type of warfare accelerates wars to an attritional character. The goal of obstructive warfare is to avoid the nearly unavoidable: that is, it seeks to deny an adversary’s ability to use “attacks from above.” This is accomplished by assailing the opponent as forward as possible and operating so prominently 2CIGI Paper No. 307 — October 2024 • Amos C. Fox within the adversary’s data networks that the opposing military force spends its time trying to make sense of data and address command decisionmaking challenges. The forward-facing attacks are not just traditional shaping — or preparatory— activities, but deliberate, planned operations oriented on confounding the adversary’s dataprocessing capability and disrupting the speed at which the adversary wants to operate. Moreover, mobile autonomous systems, which are discussed later, execute these data and tempo operations in conjunction with deliberate kinetic (i.e., military firepower) attacks. Viewed collectively, three pathways — data, tempo and kinetic — animate obstructive warfare: it uses the three pathways to defeat an adversary well before they can engage in close combat with land forces. Put another way, the goal is to obstruct a military force’s capacity to operate positionally, or to at least cause them to fight positionally in terrain that provides no tactical, operational or strategic value, and at great cost. Obstructive warfare is a theory in which states can harness the transformative potential of AI for military operations and subsequently avoid the perils of positional warfare, attrition and long and costly wars. It focuses on the functionality of new technology, and not that technology’s nomenclature or taxonomy; the purpose is to identify where the technology’s function fits within applied armed conflict. As a result, obstructive warfare is based on the assessment that systems such as drones, long-range fires, associated sensors and other like-minded technology reflect modern updates to the tools and technology that facilitate attacks from above. The phrase “attacks from above” is used throughout this paper as a noun to bypass the nomenclature of new and emerging technology and instead focus on that technology’s functionality. Depending on how one classifies attacks from above and battlefield transparency, these approaches have been a constant in war since the First World War, with the tools and methods evolving over time (Owen 2023, 26; Frontline Podcast 2022; Isbell 1993, 147). Attacks from above and battlefield transparency are relative to the technology of the day, and not a measure against future technology. This is why obstructive warfare focuses on amalgamated-system functionality instead of technologies in isolation from one another, or the euphoria of titillating dronestrike videos on social media (Rogers 2023, 73). The attacks from above strategy, however, aligns with how policy makers and senior military leaders want to use force today, which is to limit the commitment of their own land forces, yet be able to strike adversary military forces in distributed locations across the globe (Skove 2024). As US General Officers James Rainey and Laura Potter write, “A military force able to immediately link these sensors to extended-range weapons capable of precisely hitting moving targets will have a distinct advantage over any adversary” (Rainey and Potter 2023). Moreover, Rainey and Potter, among many others, insist that utilizing small, dispersed forces with limited presence on the battlefield is the way to counter attacks from above (ibid.). This method of operating — utilizing attacks from above in conjunction with capping the commitment of one’s own forces — is referred to elsewhere and in this paper as stand-off warfare (Fox 2024a; McDermott and Midgett 2021, 38–39). Dan Wright (quoted in Kosloff 2024), Christian Brose (2019), John Antal (2023a) and other leading proponents of stand-off warfare assert that an AI revolution is coming. The combination of standoff warfare and AI will allow militaries to operate with global visual transparency and at speeds that exceed human comprehension, striking remotely from almost anywhere in the world. Somewhat ironically, Rainey echoes this supposition and (inadvertently) suggests that positional warfare and stand-off warfare are the answers to battlefield transparency, stating that “when you’re maneuvering, it’s going to be to emplace fires… if it’s an Army formation, their big advantage is going to be fires: rockets, cannons, joint fires, attack helicopters” (Skove 2024). Multidomain operations doctrine, project convergence as well as the slew of other sensor, precision and long-range strikecentric concepts dominating military, academic and policy discussions, clearly demonstrate stand-off warfare’s stranglehold on the topic. Nonetheless, the wars of the twenty-first century demonstrate an alternative reality, and perhaps one that is more realistic than stand-off warfare’s vision of the future. Wars of the future will remain fights over territory. These contests for control of land will continue to be fought by armies, or at least amalgamated forces fighting on land. When attacked from the sky, military forces will seek refuge in the land — whether that be in bunkers, trenches or urban areas. Attacks from the sky are empirically proven to be less effective against land 9Obstructive Warfare: Applications and Risks for AI in Future Military Operations significant land operation to clear the occupying forces. Armies — whether they be state or non-state forces — fight land wars, regardless of how they have to get to the land war. And armies fight other armies in land wars, regardless of the presence or degree of combined arms or joint capabilities one combatant might possess over the other. In considering the logic outlined above, coupled with the ideas on stand-off warfare outlined in this paper, a handful of enduring challenges of land warfare emerge. These challenges transcend the theatre of conflict and the manner in which armies travel to the land war; that is, the challenges of land warfare are relevant in both a Russo-Ukrainian-type scenario or a ChinaTaiwan scenario. Further, these challenges are relevant regardless of whether the armies have to conduct amphibious landings from ship to shore, airborne drops from a variety of aircraft or attacks on the ground in broad-armoured thrusts across international boundaries. These challenges, primarily identified in the Russo-Ukrainian conflict, but salient in all land wars, are listed below. The list below is not presented in order of priority, but instead as a general grouping to ensure that policy makers, military practitioners and scholars remain grounded in the principles of war when state or non-state actors fight conflicts for the physical control of territory. AI helps both sides of the coin as it pertains to the challenges of land war, which is why it is important for military forces to prevent positional warfare from taking hold. These actions can be thought of as the activities of land war: → Armies must be capable of taking and/or retaking territory. – Armies must not culminate (i.e., exhaust their combat power) while taking or retaking territory. – Culmination during this phase makes the army prone to: • effective enemy counterattack; and • the inability to conduct effective exploitation and pursuit(s). → Armies must be capable of clearing enemy armies from territory. Clearing, in this instance, means physically removing a recalcitrant and hostile military force from occupied territory. → Armies must be capable of holding territory. Taking, retaking and clearing territory of hostile forces often exacts a high toll on an army, leaving it in a weakened state. Armies with small, fragile force structures experience the highest toll, and are even less likely to be able to hold on to costly gains. Resilient land forces are critical to ensuring that military forces can uphold territorial gains, whereas the tools and techniques of stand-off warfare provide only marginal returns on investment when it comes to holding territory. → Armies must be capable of protecting populations. → Armies must be capable of encircling a hostile force. This is the best way for an army to maximize the effects of joint firepower. → Armies must be capable of sealing boundaries. If armies cannot effectively seal boundaries, then they will always be prone to invasion by hostile neighbours. Resilient land forces, not missiles and drones, are the first line of defence for ensuring proper border security. In land war, verbs such as take, retake, clear, hold and encircle represent the actions one military force must take in earnest against another force, pursuant to their respective political-military objectives. Taking, retaking and clearing, for instance, involve concerted combat operations against another military force. These actions represent the direct clash of forces in a struggle between national wills, industrial bases, internal and external bases of power and the grit that each state’s armed forces can bring to bear. Similarly, holding territory and sealing boundaries are not truly terrain-oriented actions, but rather ones focused on definitively defeating a hostile force intent on removing the holding force from an important piece of terrain. While the terrain is the objective, the hostile force is the mechanism through which situational success is determined. Thus, the same variables outlined above are critical — national wills, industrial bases, internal and external bases of power and the grit of one’s military force. Protecting populations is also an action oriented toward hostile military forces, which can include having sufficient military force on the ground to interdict attacks on civilians and civilian infrastructure. Protecting populations can also include actions such as providing air 10 CIGI Paper No. 307 — October 2024 • Amos C. Fox defence and air cover to prevent, interdict and counterattack hostile attacks from above. Whereas the actions listed in the preceding paragraph are force-oriented, protecting populations is more system-oriented; that is, many of the actions in this category can include eliminating individual missiles, drones or hostile actors. Land wars, which will remain the most important type of war in the future, provide many areas in which AI can be used to maximize gains and offset losses. The potential impact of AI in military operations, primarily viewed through the lens of the logic of land wars and countering attacks from above, is explored in the following sections. The Potential Impacts of AI on Military Operations (DOTMLPF-P) Due to its transformative potential, AI can be used to help militaries thrive in land wars and overcome many of the challenges of attacks from above. Operating under the assumption that military forces in the future will be required to take, retake, clear and hold territory, seal boundaries, encircle hostile forces and protect civilian populations, it is important to explore AI’s transformative potential regarding how it can assist, or even take the lead, in military operations pertaining to land war. Because most Western militaries rely on DOTMLPF-P as a frame to design administrative and acquisitive needs in order to accomplish military missions, it is a useful model to illustrate AI’s impact on future military operations. NATO, which is concerned with the interoperability of its member states, uses DOTMLPF-I. As such, this section examines AI’s utility across the DOTMLPF-P and DOTMLPF-I spectra. Doctrine Considering the United States’ leading role in Western military thought, as well as the base upon which NATO doctrine was built, multidomain operations (MDO) is currently the foundation of contemporary Western military doctrine. As a result, many Western states and NATO members are currently reconfiguring their doctrines to align with the tenets of MDO. MDO is fundamentally a fires-centric philosophy that seeks to operate in the “attacks from above” spectrum, using long-range strike, precision fires and a network of sensors and drones to generate convergence. Convergence is the synergistic effect of military operations through the networking of sensors to detect threats, identify target locations and push the resulting data through the network to allow military commanders to employ the best weapon system against the target to generate a catastrophic impact on the targeted threat (United States Army 2022, 3–4). AI provides many areas in which MDO can be enhanced, all of which are the focus today of concept and doctrine developers looking at AI’s role in future military operations. In effect, MDO is the formal articulation of a stand-off warfare doctrine. However, it is heavily focused on attacks from above, while failing to account for the logic of land wars and the positional approach it inspires in response (Fox 2020, 8–10). This shortcoming in contemporary military thought must be addressed. While continuing to develop AI’s role in advancing convergence, Western militaries and NATO must develop strategies, concepts and doctrine that integrate AI into the activities of land war.2 Moreover, Western militaries and NATO must evolve the way in which they visualize and frame the conduct of military operations. To be sure, the “deep, close and rear” area construct will not be useful, nor contribute to the prevention of positional warfare, if states intend to unlock the potential benefits of AI.3 Western militaries and NATO must develop a doctrine that defines defeat by illustrating clear causal mechanisms and links between military activities and how those activities cause defeat. Furthermore, Western militaries and NATO must structure operations and forces to generate defeat’s associated end states. For instance, defeat can be generally defined by three conditions: when engagement with a hostile military force and denying or preventing that force from taking a positional warfare position; when one military force causes the opposing force to quit the fight; and when an opposing political leader decides to end hostilities. 2 Note: “strategies, concepts and doctrine” will be referred to simply as “doctrine” henceforth. 3 See the United States army (2022) field manual for definitions of the “deep, close and rear area” construct. 11Obstructive Warfare: Applications and Risks for AI in Future Military Operations Looking to the future, in which resilient AI and networks will replace increasingly fragile human-centric operations, Western militaries and NATO must appreciate that threats will be adaptive, with learning actors’ first intent on survival, and second on winning (Fox 2023, 5–6). Threats in an AI-dominated future will be networked and operating from intent, not direct guidance. Militaries will likely maximize the use of autonomous systems, and these systems will increase the speed at which militaries can operate in turn. Thus, time will become an increasingly important military variable, as the interval between a sensor’s identification of a target — whether that be a command post, military formation or individual combatant — and the impact of a strike on that target will likely occur at a much faster speed than in the past. And if one combatant is operating in this future AI-dominated environment, it is safe to assume that the other combatant is too. Therefore, it is important to restructure the battlespace and create military formations to accommodate this new operating environment. Western militaries and NATO must further develop doctrines that account for autonomous systems, human-machine integrated formations and traditional military forces (i.e., human-centric formations). Autonomous systems will make their biggest contribution to military operations if they are provided space to operate independently of traditional military forces. For instance, AIdriven sensors and robotic formations should be afforded battlespace ahead of human-based forces to collect information pertaining to the enemy (both actively and passively); transmit or present information to the adversary for the purpose of deception; influence the enemy toward dispositions welcoming to one’s own force; and manipulate the tempo of an opponent’s operations to support the commander’s scheme of operations. In the abstract, AI’s transformative potential exists in tapping into the hitherto underexplored fulcrum of data, tempo and warfighting as a unified method of warfare. Doing so will offset some of the risk (and fear) associated with the employment of AI-driven autonomous and semiautonomous systems with human-centric forces by clearly delineating where each type of force will be located on the battlefield and how their operations will mutually support one another. From a risk perspective, however, AI’s transformative potential decreases in what militaries call the “close fight,” or the area in which direct contact between land forces occurs. At a distance, military commanders and their staff have time to sift through information, make informed decisions and move large formations or resources from one place to the next. AI is useful in this space because it can assist the machines that collect and analyze battlefield data, while augmenting the staff that generate proposed plans and recommendations for a commander’s approval. The goal here is to create separation on the battlefield between one’s close combat forces and an attacking adversary to trigger the enemy’s culmination before close combat can occur and positional warfare sets in. In close areas, or battlespaces that lack significant amounts of geographical space between belligerents, military commanders and their staff possess very little time to make critical decisions. That lack of time may cause an actor to instinctively respond or act concerning a specific situation. In doing so, the commander and their staff forgo the benefits afforded by geographic distance and time and end up generally in a positional slugging match in which the blunt force of men, materiel and firepower point the surest path to battlefield victory. It therefore follows that in the close area, AI does not likely carry much transformative potential (see Figure 1). This does not mean that AI has no place in close combat. AI will likely be harnessed to continue to expedite the speed of tactical activity and will (theoretically) improve the tactical efficiency upon which AI-infused militaries operate. But improved speed and enhanced efficiency may not necessarily equate to transformative quality, but may instead result in only marginal improvements on contemporary methods. Moreover, it remains to be seen how faster and more efficient tactical activity will impact the general character of war: there is a distinct possibility that it might contribute to wars of attrition. In terms of doctrine, AI’s most transformative potential lies in its ability to provide military commanders and their staff with more time — time to mobilize bases of power; allocate resources to prioritized formations; move personnel, equipment and weapons into advantageous positions; and prepare the battlespace for close-area combat. AI can help transform this mode of warfare by conducting data and tempo operations, linked with a commander’s enemy force-oriented defeat pathway. 12 CIGI Paper No. 307 — October 2024 • Amos C. Fox In this way, data, tempo and defeat-focused operations should be seen as a collective that is focused on creating time and geographical space for a commander, enabling them to bring their sophisticated long-range firers and autonomous and semi-autonomous combat systems to bear on the battlefield. Within this model, AI’s transformative potential lies in underwriting the data and tempo pathways, and subsequently creating the information to improve the defeat pathway (see Figure 2). Taking this logic a step further, if appropriately integrated across DOTMLPF-P(I), and infused across the data, tempo and defeat pathways, AI might well generate a truly novel way of warfare in the twenty-first century. The novelty here is in the introduction of a holistic view of warfare that reconfigures the battlespace’s arrayal to support the data, tempo and defeat pathways, while incorporating new formations that can unlock these pathways. Moreover, addressing AI-infused doctrine in this manner allows the policy maker, strategist and practitioner to recuse themselves of the hyperbole of presentism relating to things such as drones, long-range strike and other perceived novelties of modern war. In doing so, doctrine development can avoid becoming fixated on tactical innovation, where AI’s transformative impact is limited, and instead focus on areas where AI might truly achieve transformational impacts. In addition, integrating AI-driven formations, doctrine and operations into traditional military contact zones might undercut a strategic adversary’s ability to effectively implement standoff warfare and better position military forces to appropriately address the challenges of land war. Figure 1: AI Battlespace Utility Time Close Deep Far GreatLittle Evolutionary AI AI Least Useful AI Useful Transformative AI AI Most Useful Transformative AI Proximity Source: Author. 13Obstructive Warfare: Applications and Risks for AI in Future Military Operations Organization New organizations (i.e., military formations) are critical to unlocking the transformative potential of the doctrine outlined above. Innovating military organizations to capitalize on AI’s transformative potential requires more than pairing AI-enabled unmanned combat platforms with other AI-enabled unmanned combat systems, creating AI-major generals or proliferating drone swarms. Innovation must focus on how to create time and geographical separation on the battlespace through data, tempo and combat operations far forward of one’s military force with the goal of providing military commanders with options, including creating opportunities for success and protecting their forces from hostile attack. If properly implemented and administered, AI-centric military organizations and doctrine have the potential to cripple a strategic competitor’s ability to use military force to accomplish their policy goals, while offering new methods of protection to one’s own military forces and civilian population. To that end, states must develop organizations that harness AI’s power to collect information pertaining to the enemy and the operating environment; transmit data or pictures of reality to the adversary for the purpose of deception; influence the enemy toward a welcoming disposition; and manipulate the tempo of an opponent’s operations. As noted in this paper’s section on doctrine, this philosophy of warfare must not be applied to the existing deep, close and rear area framework. Robotic formations need geographical space to operate in a manner unencumbered by friendly human-based military formations. In this situation, autonomous and semi-autonomous systems and formations can operate independently to gather information about a threat, purposefully present crafted information to that threat and use a mix of lethal fires and nonlethal attacks to influence the tempo of operations and weaken the constitution of the enemy force. Figure 2: Theory of Transformative AI Tempo Pathway Defeat Pathway Data Pathway Transformative AI 1. Doctrine 2. Battlespace arrangement 3. Military organizations 4. Policy Source: Author. 14 CIGI Paper No. 307 — October 2024 • Amos C. Fox In addition, AI can contribute to the conditions that preempt close combat. In an AI-rich battlespace, a military’s focus should be on avoiding close combat; in fact, military forces should attempt to defeat an opponent well ahead of close contact with land forces. Why the shift away from the traditional close area? Eliminating an adversary before they have the opportunity to fully deploy their forces and warfighting capabilities has several benefits. First, it preserves one’s forces, helping support their ability to arrive at an objective relatively fresh and not on the cusp of culmination. Second, by obstructing an adversary’s ability to occupy the physical battlespace, a force provides itself with greater freedom of action and reaction time, which subsequently gives it a wider range of options to address political-military matters. Finally, eliminating an adversary well ahead of the battle region using AI-driven robotic formations minimizes the inevitable death, destruction, collateral damage and civilian casualties caused by close combat. This might be achievable by applying a reconfigured battlespace and new, AI-infused military organizations operating forward of traditional close and deep areas. Rethinking the battlefield as separate regions might help in this process. Figure 3 shows four regions: the battle region, preparation region, tempo region and data region. Traditional military formations, semi-autonomous systems and human-machine integrated formations all operate in the battle region, which can be thought of as the traditional close area. AI in this space will likely be evolutionary as it will add incremental improvements to how land forces and joint forces participate in tactical military operations. The preparation region is somewhat like the traditional deep area, but as with the battle region, military forces there will also consist of semiautonomous systems, human-machine integrated formations and human-centric units. It will be used to functionally, positionally and temporally dislocate an adversary, and lure an adversary into unfavourable positions of relative weakness on the physical terrain utilizing ploys, tactics and strikes. This would be done if a military is not successful in defeating an adversary in the tempo region. In the tempo region, militaries attempt to win an emerging conflict or battle before it can expand into something far more significant or deadly. Military commanders might rely predominantly on autonomous and semi-autonomous systems and human-machine integrated robotic formations to fulfill their commander’s intent in the tempo region. Figure 3: Theory of Transformative AI Battle Region Evolutionary AISAS, HMI Systems and Human-Centric Formation Note: AS = Autonomous System; SAS = Semi-Autonomous System; HMI = Human-Machine Integrated Transformative AIAS, SAS and HMI Systems Transformative AIAutonomous Systems Prep Region Tempo Region Data Region Source: Author. 15Obstructive Warfare: Applications and Risks for AI in Future Military Operations The goal of AI-infused military operations should be to defeat an adversary state’s military without having to use one’s own close combat forces, killing an enemy attack in the proverbial crib with the appropriate doctrine, organizations and sustainment backbone. Doing so will lessen the potential number of challenges of land war in which one might have to encounter a hostile force. This is accomplished by negating that hostile force’s ability to attack by defeating or destroying four key features: their ability to understand; their means to advance; their ability to operate efficiently; and their ability to win. Protecting those same features for one’s own forces is the other side of the coin and of equal importance. The goal is to avoid doing this in the status quo manner by engaging in this fight within the battle region with close combat forces, resulting in the attritional toll of positional warfare. Aggressive, AI-governed autonomous and attritable machines, as well as autonomous and semi-autonomous formations operating far ahead of close combat forces, can blunt the offensive and shift the balance away from the necessity of fighting large-scale positional battles of attrition. If leveraged correctly, AI-enabled military formations should be used to defeat adversaries in the tempo region to sidestep the opponent’s ability to mass (whether practically or theoretically) for perilous combat in the preparation or battle region. AI-enabled robotic formations should be programmed with the intent of destroying an opposing enemy’s land or joint force where the data and tempo regions intersect. These robotic formations, free from the fear of human casualties, will allow militaries to aggressively identify and eliminate a strategic competitor’s warfighting capabilities, while simultaneously protecting one’s own forces. Forward-thinking states and their militaries should invest in mobile robotic strike groups, mobile robotic tempo groups and mobile robotic data groups to accomplish this new doctrine. To blunt hostile forces, these formations can be filled with autonomous and semi-autonomous systems, sensors, air defence systems, data transmission, formation facsimiles, generative sustainment, self-sufficient power generation and strike capabilities that operate untethered by human-based forces and at a faster pace, giving policy makers and senior military leaders struggling with how to win a specific conflict the benefit of additional time and information. Policy makers, senior military leaders and science and technology experts should also be wellgrounded in stand-off warfare, particularly in terms of how it contributes to positional warfare and accelerates wars to an attritional status as a result. These individuals do not need to be experts in the military arts and sciences, but they should possess at least a working familiarity of the causality between these various forms of warfare. In this way, they will be able to maximize the potential of AI in military operations to prevent strategic competition from drifting into wars of attrition. Policy and Interoperability As it currently stands, most Western states, led by the United States, lack a coherent and complete policy for how to use AI on the battlefield. Many are already working on this, as exemplified by the US government’s Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence (2023), and the US DoD’s Data, Analytics, and Artificial Intelligence Adoption Strategy: Accelerating Decision Advantage (2023). The US DoD is particularly focused on the ethical use of AI; in 2020, it articulated five principles for its ethical military use. Those principles are: responsible use; equitable use (i.e., the mitigation of unintended bias); traceability; reliability; and governability (The Joint Artificial Intelligence Center 2020). Nonetheless, additional work is required. States and their defence departments or ministries need to codify authority for use and responsible actors. Authority for use should clearly articulate where the decision authority resides for using the various types of AI-driven autonomous systems and where humans must be directly involved in their operation as opposed to simply overseeing it. Similarly, defence departments and ministries must also identify the responsible authority in the event that AI-driven autonomous systems conduct military activities that are immoral, unethical or illegal, or that violate the norms of international humanitarian law and the law of armed conflict. If Western states and NATO attempt to wait until those systems are fully fielded before putting these measures in place, they will have waited too long. Furthermore, interoperability cannot be overlooked. National caveats must be considered now, 16 CIGI Paper No. 307 — October 2024 • Amos C. Fox before autonomous systems are fully fielded. AI-related policy must focus on the nexus of allies and partner states in relation to laws governing the use of AI, information systems and networks. The drafting of potential caveats must be done with allies and partner states now to prevent some states from bowing out of supporting military operations at a future date. Interoperability also implies that all allies and partners possess the command and control of network systems to operate and manage autonomous systems in order for military operations to be truly multinational. Therefore, as states look to industry to develop AI-enabled autonomous systems, those systems must come with the requisite tools to allow all alliance members and partners to employ, monitor and control those systems. Otherwise, wartime command structures will not truly reflect the power of a multinational alliance, but rather only that of those states most invested in autonomous systems. Conclusion: Policy Recommendations In closing, the Russo-Ukrainian War provides a set of useful considerations for Western states and NATO. These considerations are not specific to Russia or Europe, but apply to any conflict in which a fight for territory is the goal. If China were to invade Taiwan, for instance, and NATO were required to assist Taiwan in extricating Chinese forces from the island, the challenges of land warfare outlined above would remain germane, regardless of the naval, air or contested logistics challenges also associated with that situation. Nonetheless, NATO and its members must not become blinkered by the sensationalism of stand-off warfare. Drones, long-range strikes and precision warfare all represent the continued challenges of “attacks from above,” which soldiers have addressed since the First World War. When strikes from above dominate the battlefield, soldiers go below ground; when soldiers go below ground, static battlefields develop. Positional warfare and attrition are where the costs mount in war, whether that be in the form of dead and wounded soldiers, lost resources, civilian casualties, collateral damage or even national prestige. When static battlefields develop, positional warfare replaces manoeuvres and conflicts drift into wars of attrition. NATO should consider this hypothesis: While stand-off warfare paradoxically accelerates wars of attrition, a more weighted land campaign lightly supported by joint elements better enables mobile warfare, thus unlocking a quicker and less destructive war. It therefore follows that if NATO members want to avoid wars of attrition, they should further examine this line of logic through experimentation. Wargames and tabletop exercises might suggest that stand-off warfare and “attacks from above” are the solutions to the challenges of future warfare, but they are actually causing more problems than they are solving. NATO should also take pause and examine the relationship between battlefield transparency, targeting, force design, dispersed operations and future military operations. One of the major talking points to emerge from the Russo-Ukrainian War, which is a continuation of the discourse from the 2020 Nagorno-Karabakh War, is that sensors and drone technology are obviating large land forces, making these new implements just as slow and unwieldy as tanks (McFate 2019, 231) and towed artillery (Roque 2024), both relics of a bygone era of armed conflict. Other experts who are anticipating a potential future conflict with China have made similar arguments (Underwood 2022). To address this challenge, change advocates assert that NATO forces must become smaller and lighter and operate with dispersed operations to defeat battlefield transparency, enemy drones, and threat missile and artillery targeting, among other hightechnology threats in the future (Judson 2023). As analyst Frank Hoffman (2024) contends, military operators only think through the problem of being seen by their enemy, and fail to consider the challenges that armies have to address once they have reached their objective. Thus, it would be prudent for NATO policy makers, military leaders and pundits to think through military operations from beginning to end, rather than excluding the latter at the expense of the former, which contributed to the US military’s failures in both Afghanistan and Iraq. NATO policy makers must appreciate that this approach requires resilient and robust — not light, small and disperse — land forces. NATO requires land forces that can make their way through the rigours of a transparent battlefield and array ready forces with sufficient 17Obstructive Warfare: Applications and Risks for AI in Future Military Operations combat power to meet the challenges of land warfare. Light, small and dispersed land forces fighting in stand-off warfare will not be able to defeat an ensconced challenger intent on retaining confiscated or annexed land. Strikes from the sky, regardless of how precise or deftly adjudicated, will not effectively eliminate those land forces. Ruggedized, resilient land forces — human, humanmachine integrated, robotic or otherwise— are needed to accomplish that task. Thus, NATO policy makers, military leaders and other supporters should advocate for the development of larger, more armoured land forces, while at the same time making it clear to policy makers why larger, not smaller, land forces are needed. In overcoming the challenges of land warfare — including any future war with Russia, China or even Iran or North Korea — stand-off warfare, precision strikes and long-range fires would only play small supporting roles. The majority of combat would occur on the ground between land forces, which means that the victor would have to be capable of surmounting the seven challenges of land warfare outlined within this paper. In achieving a clear victory using the methods outlined above, the winning forces would simplify diplomacy for NATO policy makers. 18 CIGI Paper No. 307 — October 2024 • Amos C. Fox Works Cited Ahern, Stephanie. 2022. “Future Landpower Conversation about Army Futures Command.” Strategic Landpower Symposium, U.S. Army War College, May 10–12. YouTube video, 37:21. www.youtube.com/watch?v=SsrLuisttr4&list=PLYtGRVo8O_ W2eLXdXigoftZ7osZjx92fU&index=1&t=1017s. Allen, Patrick D. 1995. “The Need to Represent a Wide Variety of Battle Types in Air-Ground Combat Models.” Military Operations Research 1 (3): 19–25. www.jstor.org/stable/43940683. Antal, John. 2023a. “Artificial Intelligence and the Accelerating Tempo of War.” Statement for the US Senate AI Insight Forum, December 6. www.schumer.senate.gov/imo/ media/doc/John%20Antal%20-%20Statement.pdf. ———. 2023b. Next War: Reimagining How We Fight. Havertown, PA: Casemate. Apple, R. W., Jr. 1991. “Death Stalks Desert Despite Cease-Fire.” The New York Times, March 2. https://timesmachine.nytimes.com/ timesmachine/1991/03/02/741091.html?pageNumber=6. Barnes, Julian E., Eric Schmitt, David E. Sanger and Thomas Gibbons-Neff. 2023. “U.S. and Ukraine Search for a New Strategy After Failed Counteroffensive.” The New York Times, December 11. www.nytimes.com/2023/12/11/ us/politics/us-ukraine-war-strategy.html. Bender, William J. 2024. “Air Force: Information Dominance Vision.” US Air Force. www.safcn.af.mil/Portals/64/ documents/AFD-150112-026.pdf?ver=2016-07. Barno, David and Nora Bensahel. 2022. “The Other Big Lessons That the U.S. Army Should Learn from Ukraine.” War on the Rocks, June 27. https://warontherocks.com/2022/06/the-otherbig-lessons-that-the-u-s-army-should-learn-from-ukraine/. Brose, Christian. 2019. “The New Revolution in Military Affairs: War’s Sci-Fi Future.” Foreign Affairs 98 (3): 122–34. www.foreignaffairs.com/united-states/ new-revolution-military-affairs. Cancian, Mark F. 2022. “Is the United States Running out of Weapons to Send to Ukraine?” Center for Strategic & International Studies, September 16. www.csis.org/analysis/ united-states-running-out-weapons-send-ukraine. Collins, Liam, Michael Kofman and John Spencer. 2023. “The Battle of Hostomel Airport: A Key Moment in Russia’s Defeat in Kyiv.” War on the Rocks, August 10. https://warontherocks.com/ 2023/08/the-battle-of-hostomel-airport-a-key-momentin-russias-defeat-in-kyiv/. Daniels, Jeff. 2017. “ISIS fight shows US military can use lower-cost weapons with lethal results.” CNBC, July 19. www.cnbc.com/ 2017/07/19/isis-fight-shows-us-can-use-lowercost-weapons-with-lethal-results.html. Defense Advanced Research Projects Agency. 2020. “Creating Cross-Domain Kill Webs in Real Time.” September 18. www.darpa.mil/news-events/2020-09-18a. Detsch, Jack. 2024. “Ukraine Is Still Outgunned by Russia.” Foreign Policy, April 23. https://foreignpolicy.com/2024/04/23/ ukraine-war-artillery-shortage-production-military-aid-bill/. Entous, Adam and Michael Schwirtz. 2024. “The Spy War: How the C.I.A. Secretly Helps Ukraine Fight Putin.” The New York Times, February 25. www.nytimes.com/2024/02/25/ world/europe/cia-ukraine-intelligence-russia-war.html. Fox, Amos C. 2017. “A Solution Looking for a Problem: Illuminating Misconceptions in ManeuverWarfare Doctrine.” Armor 192 (4): 17–26. ———. 2020. “Getting Multi-Domain Operations Right: Two Critical Flaws in the U.S. Army’s Multi-Domain Operations Concept.” Land Warfare Paper 133 (June): 1–12. Arlington, VA: Association of the United States Army. www.ausa.org/ sites/default/files/publications/LWP-133-Getting-MultiDomain-Operations-Right-Two-Critical-Flaws-in-theUS-Armys-Multi-Domain-Operations-Concept.pdf. ———. 2021. “Maneuvre is Dead? Understanding the Conditions and Components of Warfighting.” RUSI Journal 166 (6–7): 10–18. https://doi.org/10.1080/03071847.2022.2058601. ———. 2023. “Move, Strike, Protect: An Alternative to the Primacy of Decisiveness and the Offense or Defense Dichotomy in Military Thinking,” Landpower Essay 23-4. Association of the United States Army, June 7. www.ausa.org/publications/ move-strike-protect-alternative-primacy-decisivenessand-offense-or-defense-dichotomy. ———. 2024a. “The Principles for the Future of Warfare and Stand-Off Warfare.” Association of the United States Army Land Power Essay 24-4. April 22. www.ausa.org/ publications/principles-future-warfare-and-stand-warfare. ———. 2024b. “The Russo-Ukrainian War: A Strategic Assessment Two Years into the Conflict.” Land Warfare Paper 158. Association of the United States Army, February 20. www.ausa.org/publications/russoukrainian-war-strategic-assessment-two-years-conflict. ———. 2024c. “Setting the Record Straight on Attrition.” War on the Rocks, January 30. https://warontherocks.com/ 2024/01/setting-the-record-straight-on-attrition/.