4 December 2021

Russian Thinking on the Role of AI in Future Warfare

Anya Fink

Russian political and military leaders have increasingly pondered the challenges and implications of growing “informatization” across the armed forces for the future of warfare. A key topic of discussion has been the subject of artificial intelligence (AI). To be sure, AI has become a buzzword across the Russian government writ large, buttressed by President Putin’s personal interest in how it could contribute to Russia’s economic development.2

Highlighting the significance of AI for the Russian military, the head of the General Staff Academy Colonel General V.B. Zarudnitskiy recently wrote that an important direction for development of Russian armed forces will be “the introduction of [AI] capable of self-learning and analysis of big data for applications in various fields – from intelligence and weapons management to strategic forecasting and decision-making”. He further pointed out that “the rapid development of both military and non-military means of confrontation, primarily with the use of [AI] technologies, [is contributing to] the emergence of promising forms of employment of the Russian armed forces – from a strategic operation of general-purpose forces and an operation of strategic deterrent forces to a global military campaign”.3 In other words, the ability to harness AI-enabled technologies successfully will have an impact on the Russian armed forces and how they fight.

In Russia, while there is much breathless media coverage of specific military systems with features of AI and autonomy, more serious and comprehensive analysis remains scarce. One exception to the rule is the work of an important Russian military thinker Major General (Res.) V.M. Burenok. Burenok is the current president of Russian Academy of Rocket and Artillery Sciences (RARAN), though his intellectual home is the Ministry of Defense’s (MOD) 46th Central-Science Research Institute (TsNII), an organization which he previously headed, and with which he maintains a close relationship. Both RARAN and the 46th TsNII focus, among other MOD research topics, on analyzing approaches to the development and procurement of prospective armament systems. Burenok has long been at the forefront of significant Russian military debates: over a decade ago, he spearheaded research on the now-doctrinal concept “strategic deterrence through the use of force” related to the employment of precision strike for escalation management.4 More recently, his work has focused on approaches to developing AI-enabled systems for the Russian military.

The April 2021 issue of Military Thought features an article by Burenok focused on the “role of artificial intelligence in the military confrontation of the future”. This is not the first time Burenok has written on the subject.5 But this specific article is a primer in Military Thought, the flagship journal of the Russian General Staff, and it is an important one for Western observers. In the article, Burenok posits that “the creation and development of [AI] systems is currently becoming one of the most important areas of scientific and technological progress, the very fundamental technology that can radically change the nature of not only armed struggle, but also the whole essence of power confrontation between states, including economic, information and cyber war. This change will be characterized by the priority role of [AI] systems during this confrontation”.6 This review examines and summarises Burenok’s key points.

Automation vs intellectualization

Burenok begins by pointing out the unfortunate “fashionable tendency” in the Russian media and across Russian military writings to conflate systems with automated command and control (C2) elements or “any objects that have even insignificant levels of command programming based in uncomplicated algorithms” with AI. This is problematic, he argues, because “the inability to understand the essence of AI can lead away from the development of true AI, the complexity of the creation of which is immeasurably higher than the most complex control system”.

He then provides the familiar distinction between automation and intellectualization, arguing that it is “the realization of the computer's ability to make decisions under conditions of significant uncertainty, based on heterogeneous and incomplete information, and frequently changing situations. Self-learning and adaptability are also of great importance as the ability of the AI system to independently improve the software embedded in it, including self-programming in situations, the reaction to which is algorithmically not provided”. In turn, “we can say that AI is the ability of a computer to make decisions in various and rapidly changing situations, similar to a person”.

Burenok observes that – at present – automated systems are focused on a narrow range of tasks, and argues that it is thus reasonable to expect that AI-enabled systems will also be similarly focused because “it makes no sense to create an AI system designed to solve a wide range of tasks, like a person, it will be irrational both in terms of labor costs for its creation and in the efficiency of its application”. But key “qualities like adaptability and self-learning will remain the main distinction between AI and automated control systems”, he concludes. He then posits that the development of algorithm-based weapons systems is easier than automated C2 of forces – a significant priority for the Russian military – because the former is a much more circumscribed task. In contrast, C2 of forces is significantly challenged by incomplete information and complexity. “In the course of hostilities, similar situations do not repeat themselves, therefore, it is practically impossible to create algorithms suitable for all cases of command and control. As a result, the automation of these processes is so far only a tool for preparing the initial information for the commander to make a decision”.

Trends of AI employment in the military sphere

Burenok then spends the rest of the article discussing areas of military development where AI could be used, and discussing development priorities and challenges.

AI systems could be used in the following military areas:

Control of the technical state of the objects of infrastructure and armaments/military equipment

Forecasting and prevention of accidents at military facilities

“Autonomization” (robotization) of armaments/military equipment

Creation of robotic complexes with various levels of autonomy

Reconnaissance, monitoring, information support of forces

Collection, processing, and integration of information from various sources, formation of a complete picture of the combat zone, possibilities and character of action of the adversary, and the formation of the idea of action of own forces

C2 of forces

Development of control commands, depending on the developing environment

Forecasting

Forecasting of the military-political situation, character of possible adversary actions and the character of armed combat on the basis of processing of large volumes of data of diverse types

Combat and counter-combat in cyberspace

Cyberattacks on adversary objects, search for vulnerability in own cyber networks and the halt of adversary intrusion into them

Information warfare

Conduct of propaganda and counter propaganda in the net, information diversions, artificial bots

Logistics

Formation of plans of technical support of armaments/military equipment, plans of delivery of spare parts and components and logistical means for armed forces activities support

Construction activity

The determination of the outlook of future weapons based on the processing of volumes of technical documentation from various sources, analysis of tendencies, technical and technological achievements

From this broad picture, Burenok shifts to specific trends he sees emerging worldwide “in the interests of maintaining the combat readiness of troops and developing weapons systems”. He notes the following:

“Intelligent information systems for assessing the situation, planning and controlling combat operations in real time, ensuring the interaction of diverse forces and assets;

Multi-agent modeling for assessing the combat effectiveness of modern and advanced weapons and military equipment, weapons complexes and systems, groupings of forces and assets;

Systems for processing unstructured information on the current state and prospects of development of various fields of knowledge; assessing the impact of scientific, technical and technological advances on changing the properties of materials, structural elements, components and samples of armaments/military equipment, including electronic component base, industrial technologies for design, testing, production and ensuring the operation of armaments/military equipment”.

Burenok then posits the following priorities for AI technologies in the military sphere worldwide:

“Systems for processing and integrating information and intelligence data, including acoustic (sound and voice), optical, electronic classifications on this basis of threats and identification of targets;

Control systems for group actions of robotic, crew and mixed groupings of armaments/military equipment, including reconnaissance, shock-reconnaissance, support, including both macro- and nanorobots. Technologies of group control of robot “swarms” are being intensively developed abroad in the direction of creating promising systems of armed warfare in the air and at sea (for both surface and underwater operations). The issues of creating space systems for various purposes on the basis of mini- and nanosatellites controlled by AI are being worked out, which will ensure both an increase in the efficiency of their performance of immediate tasks and stability against the enemy;

Systems of optimal target distribution based on intelligence about the enemy (including data on his troops, weapons and their effectiveness) and an assessment of the capabilities of his troops and armaments/military equipment (AI will allow to more quickly identify and prioritize the destruction of targets, form the plans of subsequent actions of troops, flexibly respond to changing situation in real time)”.

Burenok further argues that in the not-so-distant future AI will be used for drone swarms with capabilities to destroy or disable adversary’s armaments and military equipment. And, “with the advent of quantum computers […] AI can be used in the design of new types of weapons, new materials, new designs, and even in the development of new strategies for waging war”. Indeed, “the most obvious implications of creating a truly working quantum computer is the possibility of almost instantaneous hacking of military and infrastructural encryption systems of a potential adversary, which provides tremendous opportunities in the field of both military intelligence and industrial espionage”.

Noting that there is a “new technological race” worldwide, he points out that the US and China are dedicating significant effort to developing AI technologies. Though he mentions both countries, he says nothing about Chinese AI developments and instead sticks to a brief discussion of US programs. He points out key entities and projects and notes the role of the Joint Artificial Intelligence Center as the integrator. He posits that US efforts are focused on military systems with self-learning capabilities and, more specifically, the ability to counter the adversary use of autonomy. He further notes that US efforts are also aimed at “intellectualization of cyber operations” and “propaganda and counter-propaganda” on the internet as well as the algorithm-driven gathering of information aimed at “discrediting the actions of a country, its government, party leaders, and movements”. In this discussion, Burenok avoids comparing US developments to Russian ones. To be sure, a direct comparison is very difficult in the military sphere because the countries are focused on different capabilities. But broadly, in the civilian sphere, we know that the size and scope of the Russian AI ecosystem lags behind its US and Chinese counterparts.7

Challenges to be resolved

Considering all of the above, Burenok identifies a set of five challenges that still stand between the successful creation of AI systems in the military sphere in Russia and worldwide. These challenges, listed below, appear daunting, and their resolution will take a significant amount of time and scientific effort.

“The collection, processing and analysis of the military-political situation in the world and in individual regions: the development of methods for collecting, structuring, classifying and formalizing knowledge from various problem areas (political, military, military-technical, psychological, organizational, etc.) to identify and analyze threats, develop solutions to counter them, reduce tension.

The effective and accurate modeling of the situation (decision-making processes): the study and formalization of various schemes of human reasoning on the basis of heterogeneous information for the conduct of hostilities, the creation of effective programs for the implementation of these schemes in computers.

The creation of dialogue procedures for communication in natural language, providing contact between an intelligent system and a human-specialist in the process of solving problems, including when transmitting and receiving non-formalized commands in extreme situations associated with a risk to life.

The planning and control of combat activities—the development of methods for constructing control algorithms based on knowledge about the problem area, which are stored in an intelligent system and continuously received from various and heterogeneous sources of information: reconnaissance, geodetic, topographic, meteorological, hydrographic, and so on.
The training and actualization of intelligent systems in the process of their activity, creation of means of accumulation and generalization of skills and abilities”.

Concluding the article, Burenok points to the importance of “responsibility in creation” of these systems. In this regard, “giving the opportunity to any, even the most perfect system, independently to plan and implement their behavior, to change the algorithms of actions, the designer and all of humanity are taking a very risky path, which is fraught with not yet fully realized consequences”.

In sum, while the Russian military is gradually progressing in terms developing AI-enabled and autonomous technologies, its military thinkers are grappling with the potential implications of these developments for the future of warfare. If Burenok is correct, as the nature of conflict continues to evolve, the role of these technologies and their reach across the whole of the military will only grow. This also means that serious writings about how the Russian armed forces hope to use these capabilities are likely to become much more prominent in authoritative journals like Military Thought.

No comments: