Notwithstanding mounting troubles on the battlefield in Ukraine and economic hardships at home, Russia has been pursuing AI and other emerging and disruptive technologies (EDT) with an increasing sense of urgency. Traditionally, advanced technology has been considered in Russia of critical importance for military effectiveness and strategic advantage. From Moscow’s perspective, gaining or losing ground in the contest for cutting-edge military technology and more effective weapon systems will have far-reaching consequences for warfare and, thus, for national security, sovereignty, and economy. Hence, AI development appears to have a major impact on Russia’s position and corresponding influence in the international system.

Furthermore, Russia’s sustained focus on defense AI under Vladimir Putin’s leadership has been driven by other key forces working in conjunction: first, the expectation that AI may provide a major military boost to narrow the capability gap with the West; and second, a fear that new vulnerabilities created by AI could be exploited by adversaries to undermine Russia’s security, sovereignty, and place in the international hierarchy of power. The war in Ukraine has further increased the country’s emphasis on pursuing AI. Public statements by Russian authorities, including President Putin, suggest that one of the lessons Russia has been learning is that AI-enabled weapons systems and infrastructure provide clear battlefield advantages.

This chapter begins with an examination of the Russian understanding of AI and the incentives for engaging in what is seen in Moscow as an international technology race. Subsequently, it examines how the evolving Russian approach to defense AI is reflected in a range of key strategic documents providing the framework for Russia’s AI strategy and policy. Next, it analyses how Russia goes about developing defense AI. With its traditional state-driven, top-down innovation model, Russia is an outlier among global AI contenders. However, to exploit advances achieved in the civilian sector, Russia has modified its model to partly emulate the US’s and China’s approaches.

An assessment of Russian funding for defense AI is complicated by its sensitive nature, which is rarely made public, and because AI technology is not a single product but rather a component applied in almost all the Russian military EDT programs. Hence, it comes from different financing sources and cooperation platforms involving other sectors. Still, this chapter examines the general economic trends and selected available figures pertaining to civilian AI funding that shed light on Russia’s defense research & development (R&D) environment and the relationship between declared ambitions and economic realities. Finally, this chapter examines major Russian priorities related to fielding and operating defense AI systems. Russia faces numerous systemic problems and practical limitations that this chapter considers before drawing tentative conclusions about the current state and prospects of defense AI in the country. Despite the severe impediments, Russia’s government will likely continue to prioritize AI development in selected defense applications to gain a rapid battlefield advantage.

1 Understanding Defense AI and Driving Forces

The focus on AI in Russia has significantly intensified, particularly in the early 2010s. According to the official Russian definition, AI is “technological solutions capable of mimicking human cognition and performing intellectual tasks similarly to, or better than, humans” (Presidential Decree 2019). The Russian military dictionary defines it in greater detail (Military Encyclopedic Dictionary Undated) as a network of cybernetic devices that replaces human intelligence activity; provides the foundation for building an automated control system; and is applied for the search, recognition, and analysis of information, for developing of recommendations and decisions, for automatically creating and issuing commands, and as a tool for analyzing large volumes of data. AI is seen as key to improving decision-making and complex analysis in situations where there is a high degree of uncertainty, inconsistent real-time data, and severe time pressure (Military Encyclopedic Dictionary Undated).

There is, nonetheless, a certain level of confusion about the understating of what AI is. It derives from the varying degrees of autonomy and intelligence found in different systems. Russian discourse distinguishes between automation (avtomatizatsiya), as employed in automated, remotely controlled, and semi-autonomous weapon systems, and “intellectualization” (intellektualizatsiya), which refers to the integration of machine learning and other sub-elements of AI technology (Fink 2021; Nadibaidze 2022a).

Automation and weapons systems capable of operating automatically have existed in Russia since Soviet times and are, therefore, more advanced than Russian systems integrating machine learning and other sub-elements of AI-related technology (Nadibaidze 2022a). Examples include the P-700 Granit anti-ship cruise missile, introduced in 1983, which featured AI algorithms in an on-board computer (Gazeta.ru 2021); the more precise and complex P-800 Onyx rocket; and the Don-2N radio-electronic facility in Sofrino near Moscow, designed to provide automatic detection of nuclear warheads, transmission of information to anti-missile system launchers, and offering options for action (Poroskov 2022). In contrast, the process of “intellectualization” involves critical elements of adaptability, self-learning, self-improvement, and self-programming as well as “the ability to make decisions in various and rapidly changing situations, similar to a person” (Fink 2021).

Seen from Moscow, advances in AI development may propel weapon systems towards further autonomy, ultimately replacing humans on the battlefield (Military Encyclopedic Dictionary Undated). However, the Russian government’s approach to full autonomy and to weapons and target selection without human intervention appear ambiguous. In its 2019 national AI strategy (Presidential Decree 2019), Russia highlighted the importance of developing ethical norms to govern the interaction between humans and AI. Russia recognizes the dangers in using lethal autonomous weapon systems (LAWS) and emphasizes the need for humans to stay in control. The official stance is that the loss of meaningful human control of LAWS is unacceptable. The responsibility for the use of LAWS and for any unintended consequences rests with the operator of the robotic system or programming (Belousov 2022).

However, Russia simultaneously argues that developing the criteria to define meaningful human control will be difficult without politicizing the issue. Moreover, the Russian government has systematically avoided agreeing to legally binding international instruments that would prohibit the use and development of LAWS (Hoffberger-Pippan et al. 2022; GGE 2019; Nadibaidze 2022b). One notable reason is that restrictions could hamper Russia in the global AI race (Presidential Decree 2019). In July 2022, Russia’s official representation to the UN argued that LAWS also confer several benefits. Such systems, it said, do not suffer from human weaknesses, including moral and religious attitudes or feelings of revenge, panic, exasperation, prejudice, and fear. Moreover, highly automated technologies can increase strike accuracy and thus reduce harm to civilians and civilian facilities (Belousov 2022).

While the latter argument may sound hollow given the systematic Russian targeting of civilians in Ukraine, the Russian authorities seem certain about the advantages of LAWS. Two years after the 2019 national AI strategy highlighted the importance of ethical norms to govern human-AI interaction (Presidential Decree 2019), Russia approved a national code of AI ethics. It is important to note, however, that it does not carry the weight of law as these are only recommendations. Moreover, it is designed for civilian AI systems and non-military purposes only (Novyi 2021). Indeed, defense AI development under the current Russian regime is unlikely to be significantly constrained by ethical considerations.

Two key forces have driven Russia’s pursuit of defense AI: the expectation of a major military boost and the fear of new vulnerabilities AI may create.

On the one hand, the Russian authorities expect AI can help them to narrow, if not close, the capability gap with the West that has grown larger because of extensive losses suffered in the period since Russia’s reinvasion of Ukraine in February 2022. Indeed, AI-enabled weapons systems and other EDTs promise to accelerate Russia’s military build-up and modernization in a non-linear manner, providing a military edge, if not superiority, in selected areas.

On the other hand, the Russian decision-makers are concerned about new vulnerabilities and security threats that AI can create, and that Russia’s adversaries can exploit. Some Russian officials worry that AI could help enemies win a conflict even before it officially erupts (Cheberko 2018). AI technologies are expected to have the potential to change the character of warfare and the dynamics of crisis escalation in ways that could undermine strategic balance and pose an existential threat to Russia (Sergeantov et al. 2022). Engaging in the AI race therefore appears not to be a matter of choice, but of necessity.

The Russian authorities argue that competing with the US and NATO symmetrically would take a long time due to limited Russian resources. Breakthrough technologies by contrast are seen as force enablers and multipliers that can give Russia an advantage relatively quickly (Zysk 2020). In 2021, Putin argued that AI could provide a qualitative leap to Russian weapons systems, including hypersonics, while enhancing other key capabilities such as lasers and robotics (Ria Novosti 2021; Vzglyad 2020).

The experiences Russia has gained by reinvading and occupying parts of Ukraine have added a sense of urgency to its development of a range of AI capacities, including command, control, communications and high-speed decision-making at all levels; high-precision weapons and nuclear weapons; unmanned systems for surveillance, reconnaissance, situational awareness, search and rescue, target acquisition and strike; air defenses, early warning, electronic warfare and space-based systems; logistics and manufacture; and offensive cyber and influence operations to shape the psychological domain.Footnote 1

2 Developing Defense AI

A series of strategic documents, concepts, and policy papers since the early 2010s testify to Russia’s rising interest in pushing AI forward. Many of the documents remain classified and this chapter does not aim to provide an exhaustive list. The examples below, however, shed light on the focus areas and extent of Russian AI development.

Key documents include the “Concept of development and combat use of robotic complexes for the period until 2025” (Newsru.com 2014) and the “Concept of the use of robotic systems for military purposes for the period 2030” (MoD 2014), both adopted in 2014. In 2016, Russia developed the “Strategy for scientific-technological development of the Russian Federation” (Presidential Decree 2016), which defines the “transition to advanced digital, intelligent production technologies, robotic systems, new materials and design methods, the creation of systems for big data processing, machine learning and artificial intelligence” as a priority.

Russian defense AI development has been closely connected to national economic development, as expressed in a batch of strategic policy documents released in 2018. In a follow-up, Russia created a series of national projects, such as the Digital Economy, which includes a development program for the 2021–2024 period called Artificial Intelligence (Markotkin and Chernenko 2020). To accelerate Russia’s development, innovation and investment climate, the focus has been on removing regulatory obstacles in public administration, healthcare, transport, medicine, education, construction communications, agriculture, fuel and energy and other fields (Markotkin and Chernenko 2020). In addition, Russia launched the federal Digital Technologies project, which focuses more broadly on advanced technologies such as robotics, quantum computing, virtual reality, blockchain, wireless communications and others (Petrella et al. 2021). The growing interest of the Russian authorities in AI has also been expressed in the form of conferences hosted by the Ministry of Defense (MoD) to facilitate contact between stakeholders across sectors (MoD 2018a; Voennoe Obozrenie 2018).

In connection with a Russian government order issued to state-owned companies to draft a variety of “roadmaps” for developing key technologies (including quantum computing and 5G implementation), Sberbank’s German Gref led work on the AI roadmap (MoD 2018a). Completed in October 2019, it identifies sub-technologies and transitional timing between stages of research, development, and commercialization. It also provides examples of target use cases and points out major obstacles and measures to overcome them (Ministry of Digital Development 2019).

Sberbank was also instrumental in preparing the draft of Russia’s first official AI strategy, which was finalized and signed by Putin in October 2019. This “National Strategy for the Development of Artificial Intelligence for the period until 2030” defines strategic goals in investment, R&D, infrastructure, educational and training programs, legal frameworks, and recruitment of talent to the military and security services. It also calls for streamlining actions to advance AI development across various sectors (Presidential Decree 2019).

One of the strategy’s main objectives is to significantly improve AI development in Russia by 2024 and to catch up with competitors in the field. While the document recognizes that Russia lags behind the AI front-runners, the United States and China, it argues that Russia could become a global leader in AI development and utilization (Presidential Decree 2019; Petrella et al. 2021). The strategy aims to increase the number of state and private entities engaged in technological innovation by 50% and to create high-performance export-oriented industries equipped with advanced technologies, primarily in manufacturing and agriculture (Markotkin and Chernenko 2020).

During the high-level Artificial Intelligence Journey Conference in Moscow in November 2022, President Putin noted several measures the Russian authorities plan to take to accelerate AI development. These included the development of new federal industrial robotics projects; the establishment of an AI development maturity index to evaluate practical results of AI implementation by Russian industries; and the development of sovereign cloud technologies. The ambition, he added (Putin 2022a), is to introduce AI technologies into every national development project and every state program as well as into the investment programs of Russian companies.

In July 2022, Russia adopted a concept for the development and use of weapons systems using AI though the document appears to be classified (Russiaun.ru 2022). In December 2022, to further accelerate AI development and large-scale implementation, the Ministry of Economic Development approved a roadmap titled “Development of the high-tech direction Artificial Intelligence for the period up to 2030.” To help implement this roadmap, Deputy Prime Minister Dmitry Chernyshenko signed an agreement on 16 January 2023 with 30 parties, including Sberbank, RDIF, and the Skolkovo Foundation (Kommersant 2023). Russian businesses that buy and deploy Russian-made solutions, including AI-enabled systems, were to be offered tax incentives and additional direct funding for upgrades starting in January 2023 (Putin 2022a).

An additional program to support the development of domestic import substitutes was launched in response to Western sanctions. Initial sanctions imposed after the Russian annexation of Crimea in 2014 led Russia to find some components in China (Gressel 2020). The more extensive sanctions that followed Russia’s reinvasion of Ukraine, together with a massive withdrawal of Western companies, have made the situation more precarious. While Russia is searching for technology partnerships in various parts of the world, including among ASEAN members, the Sino-Russian technology collaboration remains one of the largest and most promising for the Kremlin in the face of the restrictions. AI-related research collaboration between China and Russia has systematically expanded since 2016 and includes robotics, biotech, telecommunications, cyberspace, machine tools, and microelectronics (Petrella et al. 2021), as well as uses of outer space. In 2020, a 2-year initiative for scientific, technical, and innovation cooperation brought the two countries even closer (Konaev et al. 2021; Lee 2022; Bendett and Kania 2019).

3 Organizing Defense AI

As a part of the large-scale modernization program launched in 2008, Russia has gradually expanded its AI R&D ecosystem. While dominated by the traditional top-down approach that relies on state leadership and funding, it has also increasingly involved the civilian sector with the objective is to generate synergies and accelerate development by increasing the state’s access to resources, talent, and experience (Zysk 2020). Simultaneously, in line with a long-standing argument peddled by Putin (Putin 2012, 2013), the expectation is that defense innovation will stimulate the whole economy. Andrei Morozov, the Deputy Head for Scientific and Educational Activities at the Military Technopolis ERA, has similarly emphasized the dual-use nature of advanced military technology (Zakvasin 2019). The Head of the Department for the Development of Artificial Intelligence Technologies in the Russian Ministry of Defense, Vasilii Yelistratov, argued that while the MoD adapts civilian technologies for the army, the army also provides transfer in the opposite direction. This is increasingly important in the face of sanctions depriving Russia of solutions it once obtained from the West (Rosinform.ru 2022).

Over the years, the Russian authorities have created a large number and variety of cooperation platforms between the military and security services on the one hand and academic, industrial, commercial, and other private actors on the other. According to the state-owned Zvezda media group (Poroskov 2022), which is run by the Russian MoD, in 2022, more than 150 domestic industrial enterprises and research and educational organizations participated in joint military-civilian networks and collaborative platforms working on AI for weapons systems and combat operations. Overall, the MoD in cooperation with the Russian Academy of Sciences, financial institutions, leading research centers and universities has created and operates an R&D ecosystem that interacts with more than 1200 entities from 25 regions (MoD Undated-a).

Among the key Russian institutions with AI as an R&D priority are those known as “radical innovation centers,” “technopolises,” or “technoparks.” Their main objective is to join theory and practice by assembling scientists and experts who normally would not cross paths to accelerate progress from invention to full implementation (Zysk 2021).

The Advanced Research Foundation (Fond perspektivnykh issledovanii—FPI), created in 2012, focuses on developing new and potentially disruptive dual-use technologies, such as unmanned vehicles, including the Marker unmanned ground vehicle (UGV) and the Udar unmanned tank; autonomous systems and automated decision-making systems; superconductors (Liman); additive technology for polymetallic products (Matritsa); autonomous deep-submergence vehicles (Vityaz’-D); and ultra-thin materials (Tavolga) for improving camouflage and protection (Advanced Research Foundation Undated).

One of the most prominent Russian military AI R&D centers is the ERA Technopolis, which was inaugurated in 2018 to create EDTs to serve the Russian armed forces. In September 2019, a dedicated AI laboratory was established at ERA (Sosnitskii 2022). As Morozov has put it, AI cuts across almost all of ERA’s R&D activities, thus its development is seen as more of a means than an end (Zakvasin 2019).

ERA’s prioritized R&D fields have expanded over the years to include robotics, information security, small spacecraft, energy efficiency, pattern recognition, nanotechnology, nanomaterials, information and telecommunications systems, information technology and computer science, hydrometeorological and geophysical support, hydroacoustic object detection systems, geographic information platforms for military use, radiolocation and targeting for high-precision weapons, automated control & IT and “weapons based on new physical principles,” i.e. directed energy, radiological, genetic and electromagnetic weapons (MoD 2021b, MoD Undated-b; Zysk 2022).

There are three activity clusters at ERA: research, education, and production. Production is represented by the Kulibin Research and Production Centre and the Lomonosov Microelectronics Design Centre. Kulibin conducts experimental and small-scale accelerated production of prototypes of weapons and other military and special equipment from design samples created at ERA and by its partners. Focal points include metal processing, battery testing and development, and small spacecraft. The mini-factory has special workshops that can utilize 3D printing of plastics, photopolymer, metal, and ceramics as well as carbon processing (MoD 2021b, MoD Undated-b; Zakvasin 2019; Zysk 2022).

ERA has an extensive network of civilian partners, including engineering centers, financial development institutions and leading Russian universities and research institutes such as the renowned Kurchatov Institute and Rosatom’s research and production complex Dedal (Voennoe Obozrenie 2018; Zysk 2021). The Russian MoD also operates several other scientific and testing centers focused on AI, autonomy, and robotics to serve the needs of the armed forces and defense industry. These include the Main Research and Development Centre for Testing Robotics (Patriot-export.ru 2017; MoD 2018b) and the 46th Central Research Institute (MoD Undated-c). AI R&D is also taking place in laboratories of the Russian military-industrial complex focused on weapon systems, smart munitions, unmanned vehicles and systems, radio communications systems, machine learning of deep neural networks, VR technologies, facial recognition, big data, and others. Rostec State Corp. and many of its subsidiaries (e.g. Kalashnikov, Kamaz and Ruselectronics Holding) are among the most prominent actors (Rostec State Corp. 2019; Poroskov 2022).

Russia’s extensive R&D innovation infrastructure is coordinated by the MoD’s Main Directorate of Innovative Development (GUIR), which was created in February 2013 (MoD Undated-d, 2021a). Its objectives are to organize development, support scientific, technical and innovation programs and foster conditions favorable to the creation of advanced weapons and other military and special equipment. GUIR also monitors new technologies, both in Russia and abroad, not least those that could pose a threat to national security (MoD Undated-e). To further strengthen the Ministry of Defense’s role in the practical application of AI, a special department dedicated to the development of AI technologies was created in 2021 (Tass 2022b). The head of the department, Vasilii Yelistratov, highlighted the need for a database of relevant AI technologies to be assessed for potential recommendation onward to the defense sector (Tass 2023). Projects that pass the assessment are to be tested at ERA (Kashemirov 2020).

Commercial companies also play an important role in supporting AI development. Under the leadership of Gref and Sberbank, several of them that excel at AI development in their respective fields (VKontakte, Yandex, Mail.ru Group, MTS, Gazprom Neft and the Russian Direct Investment Fund) formed the AI Russia Alliance in November 2019. Their stated objective is to “facilitate and accelerate the development of AI in Russia for education, research and practical applications, and to foster a competitive market for AI solutions” (AI Alliance Russia Undated).

Furthermore, the Russian AI R&D infrastructure also extends to various academic institutions that have created AI centers and laboratories, such as the Neural Networks and Deep Learning Lab at the Moscow Institute of Physics and Technology, the Higher School of Economics, the Ivannikov Institute for System Programming of the Russian Academy of Sciences, the Skolkovo Institute of Science and Technology, Zhukovskii Institute, the iPavlov Conversational Intelligence and Dialogue Agents project, the National Centre for Cognitive Technologies at the Information Technologies, Mechanics and Optics University in Saint Petersburg, the National Research Nuclear University and the ITMO University (iPavlov Undated; CTII Undated; Ministry of Science and Higher Education 2019; Agit Polk 2018).

To improve implementation of the national AI strategy and activities under the federal project titled Artificial Intelligence (part of the national Digital Economy program), Russia created in September 2022 the government-affiliated National Center for the Development of Artificial Intelligence with nearly 9500 organizations from 15 sectors of the economy (Cnews.ru 2022). Its objective is to provide expert support and coordination for AI implementation, monitor key indicators of AI development, provide a platform for selecting prospective AI solutions for business, science, and government, and assist in implementing important infrastructure programs (NCRII Undated; Interfax 2022; Poroskov 2022).

4 Funding Defense AI

The Russian spending on defense AI is not publicly available. In addition, AI technology underlies most of the country’s military EDT programs, which further complicates the estimates of AI funding (Zysk 2020). The diversity of cooperative platforms involving AI R&D outside of the defense sector adds to the complexity (Zakvasin 2019). General economic and financial trends, as well as figures available for the civilian AI sector may, nonetheless, shed some light on financial conditions in the Russian defense AI R&D environment.

The AI development is predominantly funded by the state. Having the support of the top political and military leadership has been a key to overriding the traditional institutional conservatism that pervades Russia’s military organization, increase its responsiveness to policy change and open to innovation. Yet relying on preferential and centrally controlled state funding has also constrained competition, risk-taking and incentives for innovation. The projects to be pursued and funded are more likely to be chosen based on political criteria than true competitive merit (Zysk 2015). Bureaucratic red tape, widespread corruption and limited intellectual property rights are additional factors stifling innovation. R&D funding has also come under pressure due to an increasingly constrained economic environment, including periods of stagnation and low-level recession since 2014.

For instance, the budget of the Advanced Research Foundation (FPI) in 2013 was RUB3.8bn (about €90M). In 2014, it was reduced to RUB3.3bn (about €80M). The budget was then supposed to increase to RUB4.5bn annually (about €110M) for 2015–2016, yet instead it was reduced by 10%. Moreover, instead of remaining at about that level as intended in 2017–2018, it decreased from RUB3.8bn to RUB3.4bn. Ambitions to boost FPI spending significantly in the following years again came up short (Ria Novosti 2016; Nikol’skii and Bocharova 2018). Russian EDT investments amount to only a fraction of the billions of dollars invested by the United States and China (OECD Undated). For comparison, in 2022, the United States allocated about USD3.8bn (about €3.57bn) to its Defense Advanced Research Projects Agency (Zysk 2021).

In addition, the COVID-19 pandemic and Western sanctions have had a significant negative impact on Russian AI funding. In August 2020, Russian media reported (Cnews.ru 2020b) that the budget for the federal AI project to support implementation of the national AI strategy was slashed from RUB124.8bn for a 4-year period (about €1.78bn) to RUB27.7bn (about €400M). In April 2020, the Russian Ministry of Finance sought to block the budget of the national Digital Economy program (of which the AI project is a part) to redistribute funds to the reserve fund. The 2020 budgets of several other ministries and departments also declined. The Ministry of Communications, headed at that time by Konstantin Noskov, was unable to spend more than RUB26bn (about €370M) in budget funds on Digital Economy, according to Russian media. It is not clear why, but the department used only 73.3% of the budget allocated for the program in 2019 (Cnews.ru 2020a, b).

The 2019 AI roadmap’s assessment was that Russia would need to allocate RUB56.8bn (about €799M) for AI development over a 4-year period through 2024. However, the updated AI roadmap published in December 2022 called for only about RUB24.6bn (about €346M) through 2030, i.e., a 7-year period. The volume of expected extra-budgetary financing decreased even more dramatically: from RUB334bn (about €4.69bn) through 2024 to RUB111bn (about €1.73bn) through 2030. Similarly drastic reductions are seen in the expected volume of the domestic market for AI-based technology: while the 2019 AI roadmap had projected RUB160bn (about €2.25bn) by 2024, the figure given in 2022 was less than 10% of that: RUB14bn (about €218M) (Kommersant 2023).

While lobbying in Western capitals to end the sanctions imposed in 2014, Russian officials often claimed that the sanctions provided an excellent opportunity to strengthen Russia’s independence by creating domestic technological solutions (Kommersant 2022b, 2023). Likewise, in January 2023, the Ministry of Digital Development argued that the sanctions imposed in 2022 did not complicate the AI work of Russian companies, because most algorithms were public open-source projects available for download and modification. Still, the Ministry acknowledged that some developers experienced difficulties: their accounts were blocked, and there was an “ambiguous” attitude towards Russians programmers in various specialized communities. Most of those carrying out federal AI development projects in Russia are under sanctions and likely to encounter difficulties accessing sophisticated technology. Access to microprocessors in particular is a matter of concern (Zysk 2022). Another example is the decision of Nvidia to suspend sales in Russia, thus restricting access to graphics processing units used to power a host of AI products (Kommersant 2023).

To some degree, Russia has been able to evade sanctions and exploit loopholes in the exports control regime. For instance, the country has managed not only to continue but to double missile and tank production when compared to the pre-February 2022 figures (Barnes et al. 2023). Another illustration is the launching of a new supercomputer by Moscow State University in August 2023 (MSU 2023). It is to be used for training large AI models and a variety of AI and high-performance computing applications.

Yet there is no doubt about the detrimental effects of the war in Ukraine on Russia’s defense innovation environment. This includes the sharp decline in competition faced by domestic technology companies due to the departure of Western corporations and the withdrawal of their investments. In Gref’s assessment (Myl’nikov 2022), this exodus from the Russian market will cause losses to the national economy over the long run because “no Russian companies will be able to maintain the level of competition”—and—“where there is no competition, there is no innovation.”

5 Fielding and Operating Defense AI

Russia has been pursuing a wide range of programs to leverage AI technologies in the armed forces and security services. According to official sources, as of September 2022 the MoD’s Main Directorate of Innovative Development had accompanied over 500 projects for subsequent implementation, 222 of which were planned for completion and implementation in 2022 (Poroskov 2022).

Russia has been seeking to integrate AI into a range of key applications, including command, control, communications, and decision-making; unmanned vehicles for missions such as surveillance, reconnaissance, situational awareness, search and rescue, target acquisition and attack; nuclear, high-precision and other weapons systems; air defense, early warning, electronic warfare, and space-based systems; training, logistics, and manufacturing; and cyber operations and influence operations to shape the psychological domain.

The development of Command, Control, Communications, Computers, Intelligence, Surveillance and Reconnaissance (C4ISR) has long been defined as critical to gaining and maintaining information superiority (Sukhankin 2019a). The ability to selectively collect large amounts of data from the various domains and to analyze it and make rapid decisions, especially under time pressure, is seen as increasingly important to outperform the adversary in contemporary warfare. Russia’s top political and military leaders, including President Putin and Defense Minister Sergei Shoigu, have regularly highlighted the crucial value of rapid decision-making and of improving command and control (C2), communications and transmission systems (Ria Novosti 2021; Vzglyad 2020; Tass 2023; McDermott 2020b). The National Defense Management Center, established in 2014, that provides the main joint all-domain C2 structure, reportedly applies AI to support information collection, selection, analysis, and decision-making (Regnum.ru 2020; Edmonds et al. 2021). There are other examples of the Russian pursuit of the use of AI with large data sets to integrate C4ISR capabilities, such as in the ISBU command and control system (informatsionnaya sistema boevogo upravleniya), reportedly tested for the first time during the Tsentr-2019 strategic exercise (Markotkin and Chernenko 2020). It is to be designed to propose alternative courses of action based on its assessments of the situation on the ground. To speed decision-making, Russia is working to apply elements of AI in control, reconnaissance, navigation, and situational analysis technology (Sergeantov et al. 2022).

Among top Russian priorities for defense AI applications—defined as an “urgent task” (Putin 2022b)—are unmanned vehicles (UVs). Programs have gradually expanded to include more than 100 types of UVs at different stages of R&D, testing and implementation (Ria Novosti 2021; Vzglyad 2020; Tass 2023). Among conclusions Putin has drawn from the battlefield in Ukraine to date is that the most effective weapons systems are those that operate at high speed and almost automatically. He stressed the necessity of creating a wide range of AI-enabled UVs for missions such as reconnaissance, target acquisition and strike, and of being able to deploy the vehicles in various ways, including swarming and networked reconnaissance (Putin 2022b).

The Russian MoD sees the development of AI-enabled unmanned vehicles for air, ground, and sea-based missions as an important element of C4ISR. The focus is on expanding the speed, range, endurance, and scope of missions for the armed forces and other services, such as the Federal Security Service (FSB) border guard. Missions include surveillance and reconnaissance as well as air, ground, and underwater attack roles, including by kamikaze drones. ZALA Aero Group (a subsidiary of Kalashnikov) claims its KUB-LA kamikaze drone can use AI to select and engage targets. Likewise, the Lancet-3 loitering munitions used in Ukraine, are to be highly autonomous, including the ability to locate and destroy a target without human guidance and return to the operator if a target is not found (Hambling 2022a). Russia has also shown interest in counter drones AI technology, bomb identification and de-mining, anti-submarine warfare, deep-water missions involving hydroacoustics, air-defense detection, electronic warfare (EW) and situational awareness. Various developers are also working on drones for search and rescue, transportation, and logistics (Edmonds et al. 2021; McDermott 2020a, 2022; Palavenis 2022).

Russia also pursues autonomy to strengthen the credibility of its nuclear forces. Here the attention is on increasing the speed of assessment and decision-making, as well as force protection and penetration of missile defenses. This is illustrated by Putin’s “wonder weapons”, i.e., the Poseidon nuclear-powered and nuclear-capable unmanned underwater vehicle (UUV); the Burevestnik nuclear-powered and nuclear-capable cruise missile; and hypersonic weapons such as the Avangard boost-glide vehicle. AI and autonomy elements are also to be applied in the guidance systems of the Sarmat intercontinental ballistic missile and the Kinzhal air-launched ballistic missile.

Russia has also shown an interest in AI applications in the Aerospace Forces. A major objective is to disrupt or degrade communications, critical infrastructure, satellites, and other networks that the US and NATO depend on (Work and Grant 2019). For instance, the RB109-A Bylina system aims to collect large amounts of data and uses AI to prioritize and jam electronic signals. AI has also been tested for application in aircraft such as the MiG-35, the SU-35, and possibly the SU-57 to enhance their operations, including on-board information management and target recognition (Palavenis 2022; Tass 2021). Russia is reportedly studying AI-equipped autopilot systems, including in connection with domestic helicopters (Poroskov 2022). Furthermore, in focus are weapons based on “new physical principles,” such as electromagnetic, radiological, geophysical, and directed-energy weapons for missions such as countering unmanned aerial vehicles (UAVs) and satellites (Rossiiskaya Gazeta 2018). Scientists are also pursuing AI enhancement of air defense, such as the Pantsir-S (Tass 2022a).

Notably, various Russian weapons systems are to be upgraded with elements of AI, including high-precision weapons, the T-14 Armata tank (Izvestiya 2021); or the Uran-9 tracked UGV vehicle and the Nerekhta reconnaissance UGV that are to be introduced as a part of the Russian ground forces to carry out “experimental military service” (Cranny-Evans 2021). AI is also a component in combat robots, for instance, a fighting vehicle based on the BMP-3 infantry vehicle and Sinitsa remote-controlled combat module (Argumenty i Fakty 2022). Russia is working as well on AI projects to improve control of artillery targeting and on new-generation infantry combat systems such as the Ratnik, whose advanced elements include software linked to small UAVs and other AI-enabled systems (Sukhankin 2019b).

AI is also seen as a critical capability in offensive cyber operations, as well as cybersecurity and cryptography. The objective is to strengthen Russian information security and leverage AI on a broader scale to enhance cyber capabilities. The list of known Russian offensive cyber operations is extensive (CISA Undated; Hakala and Melnychuk 2021), including in Ukraine. Despite initial claims that Russia failed to launch cyberattacks during its invasion, research findings indicate that they have figured prominently alongside the invasion. Only in the first 5 weeks of the war, Russia conducted an intensive campaign in the cyber domain, with some 800 attacks against Ukrainian targets. The impact of Russian cyber operations has been minimized through a combination of Ukrainian preparedness and support provided by the U.S. Cyber National Mission Force, which arrived before the invasion, together with the international cooperative cyber defense task force (Corera 2022).

Furthermore, AI-enabled systems figure prominently as a tool for creating new opportunities and augment traditional methods of influence, including in disinformation, demoralization, and propaganda both abroad and domestically. One example is ERA’s project (Zakvasin 2019): a search and rescue drone with an onboard AI-enabled system capable of analyzing situations and recognizing persons that “pose a threat to society,” such as “terrorists.” Technologies such as facial and pattern recognition enhance information collection, assessment, and prediction to more effectively influence a population’s behavior. Notably, the 2022 AI roadmap published by the Ministry of Digital Development requires the regions to collect large, anonymized data sets for the purpose of training AI systems (Kommersant 2022a).

It is important to note that AI applications are also being explored to heighten productivity in the defense industry. Some production lines reportedly feature applications capable of recognizing details, tools, and human action. The objective is to reduce the role of the human factor in manual operations in the production of weapons and military equipment, such as the production of rocket engines (Poroskov 2022). The Russian United Aircraft Corporation plans to use an AI-based digital system to automatically quality-control aircraft parts for MiG fighter jets (Poroskov 2022). Likewise, state-owned Rostec has been testing the Zyfra Industrial Internet of Things Platform, which uses AI to track the engine manufacturing process and conduct simulated testing in a virtual environment. The objective is to reduce the number of tests, improve quality, and accelerate production (McDermott 2020a).

6 Training for Defense AI

The Russian authorities argue (Presidential Decree 2019) that the country’s strong intellectual traditions and high level of education in science, technology, engineering, and mathematics (STEM) will help it join the club of global AI leaders. In reality, while Russia was ranked fourth in the OECD’s 2019 global index of education, less than 1% of Russia’s graduates earned an IT, communications, or other technology-based degree. The Lomonosov Moscow State University—considered Russia’s highest-ranked computer science research institution—was listed 43rd globally in 2017, 60th in 2018, and 78th in 2019 (Dear 2019). Overall, Russia ranked 47th in the 2022 Global Innovation Index (GII 2022).

The 2019 national AI strategy highlighted the importance of education and training in AI (Dear 2019). To improve the pool of specialists in new technologies, Russia tests various strategies to train and retain a new generation of specialists. AI centers offering professional education have been established at the top Russian universities and research institutes. Many offer participation in real development projects by corporate partners, such as Gazprom Neft, MTS, Sberbank, Russian Railways, and others (Ministry of Science and Higher Education 2019; MoD 2018a). A partnership agreement was signed by FPI and the Ministry of Science and Higher Education to facilitate the creation of new scientific schools and centers of expertise focused on EDTs (Tass 2019).

The Russian authorities organize a variety of events that aim to attract university students and even schoolchildren. More than 3000 students enrolled in AI master’s programs in 2022. Medical doctors, teachers, and lawyers as well as employees in manufacturing, communications and transport can take a special AI educational module to improve their qualifications. In November 2022, to ensure training quality, President Putin ordered the ranking of universities in the AI field. He also highlighted (Putin 2022a) the need to introduce elements of AI into mathematics and computer science curricula. In October 2022, more than 19,000 schoolteachers from different regions of Russia took part in an online AI course (Ria Novosti 2022). The Ministry of Defense also organizes high-level conferences devoted to setting the AI agenda and develops AI training programs in a wargame style with tactical, operational, and strategic levels to illustrate the effect of AI on warfare and stimulate further development (MoD 2018a). Still, the Russian war in Ukraine has exacerbated long-standing problems with shortages of professional expertise. An exodus of qualified scientific personnel, including IT specialists, accelerated after Russia announced a mobilization of reservists in September 2022 (Kommersant 2023; Metz and Satariano 2022; Washington Post 2022).

Known measures to recruit and retain talent in the armed forces include the creation of “military scientific units” (nauchnye roty). Staffed by conscripts, these units have developed since 2013 on the foundation of Russian military research and higher educational institutions (MoD 2016). The MoD has gradually been transferring the units to ERA (MoD 2016), and in 2022 eight of themFootnote 2 were operating there in various R&D priority fields (MoD Undated-f; Zysk 2021), and supporting needs of several units such as the Aerospace Forces and the 12th Main Directorate. The expectation is that conscripts in these units will continue military-scientific careers working at ERA when their service period ends, either as civilian specialists or with the rank of lieutenant.

To address the problem of brain drain, the Russian authorities also resort to decrees and resolutions as well as the introduction of certain privileges and incentivized funding for academic and scientific institutions, state support for the purchase of domestic replacements of foreign technology, labor market incentives, and streamlining of procedures to employ foreigners (CNA 2022). However, with the deteriorating socioeconomic situation and increasingly repressive authoritarian rule, it is unlikely that bureaucratic measures will be sufficient to make a significant difference.

7 Conclusion

The overall development of Russian defense AI appears to be in the early stages of maturity. The primary focus is on incremental evolution: upgrading legacy systems—nuclear, strategic non-nuclear, and non-military methods and means of warfare—with new technologies. AI is being tested and used in data analysis and decision support, loitering munitions, electronic warfare, communications analysis, cyber warfare and information confrontation, to name but a few applications. Simultaneously, Russia is experimenting with “risky projects”, i.e., novel systems, materials and approaches to warfare that could potentially yield significant battlefield advantage—if not superiority—in selected areas.

Still, the high-tech development in Russia has been undermined by extended periods of economic stagnation and recession, aggravated by the COVID-19 pandemic, sanctions and a massive outflow of international corporations halted cooperation. The poor investment climate is further undermined by long-standing unfavourable demographic trends and weak educational foundations. The full impact is yet to come, but dramatic spending cuts on AI in the civilian sector have already occurred. Beyond funding, several other factors will influence the future of Russian AI. One is the extent of Russia’s ability to continue circumventing sanctions and moderating its dependence on Western technology. While Russia is reluctant to create new dependencies that can turn into a source of vulnerability, the Kremlin has little choice but to supplement AI development efforts with foreign technology, including drones purchased from Iran and electronics and various other dual-use technologies from China (Kuo 2022; Lo 2023). The pervasive structural problems plaguing Russia’s defense sector and the political and economic system at large are an additional factor hobbling AI development and innovation. Systemic reforms will be required to buoy the competitive research environment but are unlikely under the current regime.

All the same, Russia’s 2024 state budget clearly demonstrates that the Kremlin is willing to prioritize the defense sector. Despite the deteriorating national economic environment, Russia plans to increase defense spending by 25% in the 2024–2026 period (AP 2023; Wiśniewska 2022). Extensive failures during the Russia’s full-scale 2022 invasion of Ukraine have prompted a major reassessment and reforms in the Russian armed forces. How much attention will be paid to R&D in that reckoning remains to be seen. To date, the combination of optimism about significant advantages AI and fear of strengths it can provide adversaries is likely to keep Russia’s attention on selected AI applications, not least given Putin’s personal interest in this development. Indeed, the experiences from Ukraine have encouraged Russia to double down on its AI commitment. Because Russia’s constrained socioeconomic and industrial circumstances make a swift military build-up in linear fashion harder, AI development in selected areas may be the best hope to rapidly gain advantage.

To take full advantage of AI, Russia must not only harness the technology but also adapt doctrines, concepts, force structures, and recruitment patterns accordingly. The conflict in Ukraine has exposed a high degree of institutional conservatism in the Russian military. There are, nonetheless, clear patterns of Russia’s ability to learn and adapt, however slow in the initial phase of the war (Konaev and Daniels 2023). The extent to which Russia’s leadership will be able to draw the right conclusions and increase the military organization’s ability to change amid an ongoing war is yet to be seen.

Important to the US, NATO and EU countries is that major Russian weapons programs aim to either match or undermine key Western military capabilities. Simultaneously, Russia is investing in a range of AI-supported indirect and non-miliary methods and means of warfare, including offensive cyber and influence operations to undermine or bypass opponents’ strengths and exploit their vulnerabilities. These focus areas are poised to gain importance, especially during the interim period, as the militarily weakened Russia is rebuilding its armed forces.