Russian AI-Enabled Combat: Coming to a City Near You?
Being a Russia scholar interested in how artificial intelligence (AI) will influence international security and warfare, you often feel like the Jan Brady of the Bunch. With all eyes on China (Marcia! Marcia! Marcia!), Russia’s relatively limited budget for AI research and development, and its smaller private sector innovation ecosystem, seem less formidable by comparison. But when it comes to military applications of artificial intelligence, overlooking Russia is a mistake. Moreover, as top U.S. military leaders now believe that American forces will increasingly be drawn into urban combat, Russia’s investments in artificial intelligence and other emerging technologies that can help its soldiers counter the physical, cognitive, and operational challenges of urban warfare and perform better in future conflicts deserve closer attention.
Since the mid-1990s, Russia has deployed troops to urban areas in Chechnya, Dagestan, Georgia, Ukraine, and Syria for conventional offensive operations, counter-insurgency, and counter-terrorism missions. While the Russian government, its military, and its people are well familiar with the heavy toll urban warfare exerts in manpower, resources, and political capital, Russian strategists also realize that much of the fighting in future conflicts will take place in cities.
Broadly speaking, developments in military robotics, autonomy, machine learning, and artificial intelligence that improve intelligence collection and analysis, facilitate navigation and maneuver in dangerous terrain, and allow for more precise targeting can help reduce the costs of urban warfare and enhance combat effectiveness across the spectrum of military operations. Since the start of its modernization reforms in 2008, and especially over the course of its involvement in Syria, the Russian military has achieved major breakthroughs in developing a wide range of unmanned systems and in further refinement of its command, control, computers, communications, intelligence, surveillance and reconnaissance (C4ISR) capabilities — technologies that Moscow expects will safeguard soldiers’ lives and make its forces more precise and lethal in combat.
Below, we assess Russia’s current and potential technological advances, with an emphasis on AI applications, in two areas of proven relevance to urban warfare: robotics/autonomous systems and information operations.
Robotics and Autonomous Systems
The Russian military today views AI as an enabler in its automated control and decision-support systems that enables rapid analysis of diverse data in multiple domains. In the near future, Defense Ministry experts hope to develop AI capable of operations approximating human brain function. The incorporation of AI into semi-autonomous and autonomous ground and air vehicles can potentially help improve force protection, increase situational awareness, and ensure freedom of maneuver and movement in complex urban terrain.
The Russian Defense Ministry is seeing the patterns of modern warfare in the Syrian conflict and is responding to them by investing in certain technologies and adjusting tactics. Syria presents the “contours of future war” — the kind of combat that involves unmanned systems, military robotics, precision-guided munitions, and robust C4ISR as well as information operations — to none other than Chief of Staff Gen. Valery Gerasimov. Taking Moscow’s weapons already in the field under consideration — especially those used in Syria — as well as those under development, the following technologies have either already been tested in urban conditions or may be applicable for urban combat.
Unmanned Ground Vehicles
In Syria, Russian forces tested and employed a broad range of unmanned ground vehicles to perform a variety of tasks including demining; intelligence, surveillance and reconnaissance (ISR); logistics; and combat missions. Facing treacherous conditions where mines, improvised explosive devices, and unexploded ordnance endangered its personnel and allied forces, the Defense Ministry successfully tested and utilized Scarab and Sphera small unmanned vehicles for ISR, as well as the larger remote-controlled mine-clearing Uran-6 unmanned vehicle. These ground robots were reportedly accepted into service in 2018, and the Defense Ministry is now incorporating the lessons learned from their use into the official concept of operations. Concurrently, the Russian defense sector is actually working on an unmanned ground vehicle designed specifically to withstand tough urban combat conditions. The project called Shturm (Storm) is based on the T-72 tank chassis and features specific defensive technologies and offensive armaments for a city fight. Moreover, the smaller Nerehta unmanned vehicle under development is supposedly a test-bed for military AI applications.
The Russian military expects that in future wars, unmanned ground combat vehicles will accompany regular forces. To prepare for this, the Russian military revealed in May 2018 that it had tested the Uran-9 – an armored unmanned ground vehicle the size of a small tank. But only a month later, defense officials publicly acknowledged that this vehicle failed in its first “near-urban combat” mission in Syria.
The military learned that in the near future, such technology should only be used in breaching operations and in attacking stationary targets. It also acknowledged that Uran-9’s operator had trouble understanding the battlefield situation via sensor data generated by the vehicle. Such conclusions point to the need for a “smart” system capable of independent combat analysis and orientation in a difficult environment — perhaps something that a limited AI application can do. Nonetheless, the Defense Ministry claims to have solved major issues that plagued this vehicle in Syria and is supposedly accepting it into service as of January 2019.
The Russian military has invested in electronic warfare technologies over the past two decades, and the results are bearing fruit. In Syria and Ukraine, Russia may have used the Leer-3 system — it consists of a mid-range unmanned aerial vehicle that jams enemy cell towers and hijacks the cellular signal. The system is designed to send text, images, and video on “captured” phones — up to 6,000 subscribers at a time. If this is scaled up and there is a fleet of such unmanned aerial vehicles deployed in contested environments such as cities and urban centers, Russia can gain a powerful information operations tool.
In Syria, Russia used a fleet of unmanned aerial vehicles with great success not only to truly see the battlefield for the first time but also to observe enemies in nighttime and to correct artillery salvos. To the Defense Ministry, such technology use was unprecedented, and there is every indication that the military will build on this Syrian experience. At the same time, the military today is experimenting with concepts like Glaz (The Eye) — a small camera and sensor package users launch from a flare gun, transmitting data to soldiers while it descends to the ground in a parachute. Russian soldiers presumably tested Glaz in Syria, and its very design lends it to be used in urban combat.
Russia is experimenting with augmenting its ground soldier capabilities with exoskeletons. As announced by Russian military news outlets, the Ratnik-3 and Ratnik-4 soldier combat gear will feature active exoskeletons that will allow troops to carry more weapons and equipment, including personal unmanned aerial and ground vehicles, for better situational awareness. There are rumors that AI could power all of the above to help soldiers make better decisions by analyzing the information they generate and gather from other soldiers and weapons.
How, therefore, to use all of the above in the most optimal fashion, especially in complex urban environments? The Defense Ministry may have several answers. First, it recently announced that it began research and formulation of how its autonomous ground systems can function better in “conditions of ambiguity” powered by an “intelligent decision-making system.” Second, the Defense Ministry stated that it is connecting all the national early warning radars as well as anti-aircraft and anti-missile systems like S-300 and S-400 into a single “armored fist” — and that AI will assist in analyzing information all these systems receive for much faster decision-making in times of crisis. Additionally, the Defense Ministry announced that an AI-enabled automated control system will connect offensive and defensive airborne and stationary assets in Crimea — aircraft, helicopters, missile batteries, and other systems — for optimal data analysis and decision-making. Moreover, Russia’s own DARPA-like agency (the Advanced Research Foundation) announced that it will design an AI-enabled system to sift through vast quantities of satellite imagery for faster and more accurate analysis.
These concepts are still going through trials, but they already show how the Russian defense community is thinking through the integration of different systems in combat. Applied in urban settings, an AI-enabled control system that can direct and synchronize Russian soldiers, unmanned aerial and ground vehicles, electronic warfare and other assets, coupled with their growing C4ISR capabilities, could give Russian forces an advantage. This, however, hinges on secure communications to connect these systems to soldiers and command centers. Moreover, major logistical challenges exist today with using unmanned systems in combat — issues like manned-unmanned teaming, functioning in a communications-challenged environment, resupply, rearmament, and others.
Of course, the Russian defense establishment may prefer not to scale up their fight in urban conditions in the first place, and instead to send small Special Forces contingents and rely on willing allies or proxy forces. This was certainly the case in Syria, where allied militias or private contractors carried out the heaviest fighting on the ground — including in cities and towns. Still, the potential exists for Russia to change how it could fight in cities by using the afore-mentioned technologies and concepts.
Russia considers information warfare as a central tenet of contemporary conflicts and a useful strategy for advancing its foreign policy goals. Its approach to information warfare is by now well-known and relies on information manipulation; active measures (overt and covert influence operations); maskirovka (obfuscation and deception on and off the battlefield); and other uses of propaganda, misinformation, and disinformation to foment discord, confusion, and fear in the target audience. Harnessing the power of AI, as former Deputy Defense Minister Yuri Borisov stated in March 2018, will allow Russia to more effectively contest the information environment and win in cyber wars.
In the Conflict Zone
Today, any individual with a smart phone and access to the Internet can share real-time images, videos, and other content that can reach and influence millions of people across the globe. In this new information ecosystem, information operations and public affairs functions — already a challenging endeavor in complex, ethnically diverse, and socio-politically intricate urban environments — have become increasingly important for the success of the mission.
In Syria, Russia has proven quite skillful in manipulating this new information environment to its advantage. According to a senior British Army general, Russian information operations became “extremely aggressive” as the fight against ISIL wound down. Distorted narratives crediting the Assad regime for protecting civilians and criticizing the United States for supporting terrorists in Syria spread on social media as well as in Russian outlets such as RT and Sputnik in an effort to discredit the U.S.-led coalition and shape the strategic environment in Russia’s favor. Not known for its population-centric counter-insurgency tactics, Russia nonetheless considers its “information work with the population” during the operations in Aleppo critical for “liberating entire neighborhoods without a fight” and allowing more than 130,000 civilians to exit the area. Overall, they believe information operations have been extremely effective in advancing military results in Aleppo, Palmyra, Deir al-Zour, and Ghouta.
With rapid advances in AI and the increase in the share of conflicts fought in cities, we are likely not far from witnessing AI-enabled information operations in urban environments. Russia’s civilian developers are already working on facial recognition and speech recognition technologies, which can potentially serve military roles, including for information operations in urban settings. One of Russia’s most innovative facial recognition AI startups, NtechLab, even won the IARPA’S 2017 Face Recognition Prize Challenge for its FindFace application that developers originally designed to sift through profiles and recognize people in photos on Russia’s popular social media network, VKontakte. Now, NtechLab is one of the top contenders to implement AI-enabled facial recognition programs as part of Moscow’s extensive urban surveillance network, allowing city authorities and domestic security agencies to monitor the daily movements of almost 12 million people. Using similar technology in an urban conflict — powered by the data unmanned systems described above can gather — could enhance the Russian military’s situational awareness when augmented by a data-crunching AI algorithm.
In the broader context of Russian disinformation campaigns, U.S. policymakers are especially concerned about developments in machine learning propelling breakthroughs in deep-fake technology — highly realistic and difficult to detect photo, audio, and video forgeries. In May, AI engineers from the Samsung AI lab in Moscow published a paper introducing a new style of deep-fake technology that can automatically create realistic animations of a person’s face from a single photo — a marked improvement on previous models that require lengthy meta-learning on a large dataset of videos. Looking to the future and thinking about the application of this technology in urban warfare, it is easy to imagine how a credible deep-fake video depicting U.S. forces committing atrocities, for instance, can incite local unrest and fuel international condemnation. Military doctrine dictates a swift response to adversary actions to prevent them from gaining advantage in the media-saturated urban information environment. But even if the deep-fake video is debunked, recent research shows that false news spreads “farther, faster, deeper, and more broadly.” In urban operations, the spread of disinformation at machine speed can then pose a serious threat to the troops on the ground and the viability of a mission as a whole.
Targeting the Enemy’s Home Front
In addition to trying to control and shape the information environment in the conflict zone, information operations also often target the adversary’s domestic audience in an effort to weaken political support for the mission. This is particularly important in the context of urban warfare because urban operations are resource- and manpower-intensive, time-consuming, and typically responsible for high military and civilian casualties. The “battle of narratives” surrounding urban operations can become highly contested and consequential, and successful information operations may even be able to shift the balance of forces or the balance of resolve on the battlefield itself.
In at least one instance, Russian information operations in Ukraine targeted the families of soldiers deployed to the war in the East. The families received a text saying, “Your son is killed in action,” prompting them to call or text the soldiers. Shortly after, the soldiers got a text urging them to “retreat and live,” followed by an artillery strike on the location where the Russian-backed separatists detected a large uptake in cellphone activity. Such coordination demonstrates Russia’s ability to combine electronic warfare, cyberwarfare, information operations, and kinetic force to hurt its adversary not only on the battlefield but also on the home front. Considering evidence of “significant and persistent interactions between current and former [U.S.] military personnel and a broad network of Russia-focused accounts” on Twitter, it is possible that future information operations will target active and retired U.S. military members.
It should be clear by now that Russia aims to master the psychological dimension of information operations, focused on undermining the state institutions and belief systems of its adversaries. And technological advances in AI have the potential to “hyperpower Russia’s use of disinformation.” It’s well known that Russian influence operations and election interference campaigns leveraged machine learning to tailor propaganda to specific audiences based on race, ethnicity, ideology, demographics, and geographic location. Yet, while the thousands of fake and stolen social media accounts deployed during these operations were manufactured manually, AI can be used to automate, accelerate, and scale synthetic accounts and content. Another worrisome example of an emerging technology the Russian defense establishment could potentially use in information operations entails recent developments in text-generation technology that mimics how humans write. Experts are concerned that this technology will allow governments and non-state actors to spread disinformation on a tremendous scale, as well as that disinformation campaigns can evade detection by generating subtly different content. In this sense, developments in AI could make Russian information operations more efficient, far-reaching, and widespread.
Russia has a long history of waging information warfare. But modern technology not only furthers its reach and magnifies its impact but also arguably renders its democratic adversaries — with their open societies and free flows of information — more vulnerable than ever before. Looking ahead, Russia is likely to continue leveraging machine learning and advances in natural language processing to refine its micro-targeting of malicious content and to construct emotionally sophisticated and relevant propaganda for more effective information operations in future conflicts, including, and especially, those conflicts that entangle Western forces in urban fighting.
Predicting how AI will impact the future of strategic competition and warfare is difficult because it requires us to assess technologies that are still mostly immature. That said, contextualizing militarily relevant AI applications and other emerging technologies within the appropriate operations environment is the best way to understand their potential impact on the battlefield.
More extensive AI and autonomous capabilities infused into Russian armed drones and unmanned ground vehicles, as well as the incorporation of AI as an enabler of rapid command, could potentially undermine the U.S. military’s ability to maintain overmatch in multi-domain battle. Overall, Russian advances in military applications of AI threaten to erode American technological and operational advantages on future battlefields, including in urban warfare.
Russia is also likely to capitalize on breakthroughs in AI, big-data analytics, and machine learning to conduct more targeted, scalable, and impactful information operations, which should alarm both civilian and military U.S. decision-makers. Thus far, the bulk of commercial and defense investments in the application of AI for detecting, analyzing, and countering disinformation have largely focused on identifying and filtering out malicious content and blocking bots. These are, at best, damage control measures. Moreover, such an approach is inherently limited given the speed within which disinformation and propaganda spread, and the significant and often irreversible damage such disinformation can cause for public opinion and perceptions of legitimacy in urban military operations.
Russia will probably not lead the world in AI in the near future. But underestimating its ability to leverage advances in AI and other emerging technologies to further its broader strategic goals would be unwise.
Dr. Margarita Konaev is a research fellow at Georgetown University’s Center for Security and Emerging Technology interested in urban warfare, military applications of artificial intelligence, and Russian military innovation in emerging technologies.
Mr. Samuel Bendett is an advisor with the Russia Studies Program at CNA where his work involves Russian defense and security technology, unmanned military systems, and artificial intelligence development. At CNA, he is a member of the Center for Autonomy and AI. Bendett is also a fellow in Russia Studies at the American Foreign Policy Council.
Image: Russian Ministry of Defense, modified