Before Russia’s full-scale invasion in February 2022, Ukraine had an advanced ecosystem to produce artificial intelligence (AI) solutions for commercial use. The country also ranked first in the number of AI companies in Eastern Europe (Oxford Insight 2020) with more than 7000 AI engineers and hosting research and development (R&D) offices of multinational companies like Amazon, Lyft, Google, Samsung, Grammarly and many others. In addition, the country approved its first national AI strategy (Cabinet of Ministers 2020) in 2020, which, among other priorities, also identified the importance of implementing AI in the field of security and defense. However, Ukraine’s defense sector only used AI sporadically, as the defense industry was primarily state-owned, focused on traditional hardware, and was not known as a “hotbed” for innovative, especially software-driven solutions.

The current war between Russia and Ukraine has changed almost all these parameters. Unlike other conflicts over the last 30 years, this is a war among technologically advanced countries. Therefore, both sides began developing and fielding AI solutions for tasks like geospatial intelligence, operations with unmanned systems, military training, and cyber warfare. Consequently, the war in Ukraine became the first conflict where both parties compete in and with AI, which has become a critical component of success.

In 2022, Ukraine’s resilience originated from a more active use of awareness systems, volunteer and private AI initiatives as well as the crowdsourcing of data, the use of open-source platforms, and advances in the decentralized use of the respective assets. As of 2023, the Cabinet of Ministers began to launch policy initiatives more systematically to advance defense AI, to advance the respective national industrial base, creating a new ecosystem, and advancing public and private funding solutions. The ecosystem of information technology (IT) companies also began to shift its focus from developing civil products to dual-use products and actively recruit the best talents in the AI field. Foreign AI solution developers are using the war as a testbed to evaluate the performance of their solutions on the battlefield. Some defense observers even contend that AI solutions that have not been tested for use in Ukraine’s combat environment won’t be fit for purpose.

This war poses challenges not only to AI solutions providers but also to AI regulation as battlefield experience prompts the need to reconsider the appropriateness of the “human in the loop” principles and other core assumptions. It seems that many of the existing regulatory principles will be hard to sustain as they have not been designed for war and thus hamper countries in their need to defend against aggressors. Consequently, Ukraine is turning into a most valuable test case for the development and regulation of war-proof defense AI solutions.

1 Thinking About Defense AI

At the time of writing this chapter, Ukraine is at war. Almost all investments, intellectual and material resources are directed toward survival and enhancing defense capabilities to ensure success on today’s battlefield, not in 2 or 3 years.

This focus also drives defense AI in Ukraine. According to the national AI strategy, AI is understood as

an organized set of information technologies, with the use of which it is possible to perform difficult complex tasks by using a system of scientific research methods and algorithms for processing information obtained or independently created during work, as well as to create and use own knowledge bases, decision-making models, algorithms for working with information and determine ways to achieve the tasks (Cabinet of Ministers 2020).

Ukraine’s current efforts are geared towards using AI to establish a competitive edge vis-à-vis Russia and build an ecosystem that will sustain this technological advantage to withstand aggression and rapidly develop much-required defense solutions. While both sides benefit from using defense AI to advance the processing of information, support decision-making, and enhancing existing equipment with defense AI, the question is how to get the upper hand in this technology-driven “tit for tat” game. In this regard, the accumulation of knowledge and experience could become a future strategic currency for Ukraine to advance national force development and share insights with allies and partners. The battle use of defense AI also yields valuable lessons for the future regulation of defense AI. Considering the situation, Ukraine has started to explore how warfighting insights on defense AI could be transitioned from the current state of war into a future state of peace as it is particularly important that data, which has been collected during wartime, would henceforth not be used in ways detrimental to human rights and democratic values.

1.1 Regulating Defense AI

The Ministry of Digital Transformation of Ukraine (Cabinet of Ministers 2023) is the main body responsible for policies in the field of AI and robotics development. Within the Ministry, a separate department has been established, and since 2020, there is an expert committee on developing and deploying AI.

So far, the national AI strategy (Cabinet of Ministers 2020) is the country’s capstone document, which also foresaw the use of AI for national security and defense as a key area of application. According to the strategy, AI solutions must ensure the protection of human rights and adhere to ethical standards. To this purpose, the strategy also includes developing an ethical code on AI but work in this direction has not yet started.

National security and defense are key use cases for AI in Ukraine. According to the strategy (Cabinet of Ministers 2020), the government wants to develop solutions for the use of AI for

  • Command and control (C2);

  • Weapons and military equipment including unmanned systems and unmanned demining solutions;

  • Collection and analysis of information during combat operations;

  • Analysis in support of intelligence activities including the processing of cartographic information;

  • Defense against cyber threats including applications that allow for the quick detection of cyberattacks, preliminary scanning, and subsequent avoidance of malicious codes;

  • Simulation and cognitive modeling of the combat situation.

These use cases suggest that Ukraine sees AI as integral to its attempt to develop unmanned systems and advance cyber defense solutions as envisioned by the country’s 2021 Military Security Strategy (President 2021). To this end, the Cabinet of Ministers has also adopted an implementation plan (Cabinet of Ministers 2021). The plan outlines a series of initiatives such as adopting legal regulation on issues of forming state policy in the field of AI, providing state support to use AI in priority sectors of the economy, implementing AI to enhance national cybersecurity systems, and defining priority tasks and directions for the current and future use of AI for defense.

In addition, Ukraine participates in international initiatives to regulate the development and use of AI. In particular, the Ukrainian Parliament ratified the agreement with the European Union (EU) on joining the “Digital Europe” program in February 2023 (Verkhovna Rada 2023) and signed the November 2023 Bletchley Declaration on AI Security at the AI Safety Summit (Ministry of Digital Transformation 2023a). As an official candidate for EU membership, Ukraine clearly signals its will to synchronize legislative efforts to advance digitalization with the EU’s policy initiatives. But while the Ministry of Digital Transformation has outlined a roadmap for regulating the civilian and commercial use of AI in Ukraine (Ministry of Digital Transformation 2023b), the respective capstone document for the use of AI in support of national security and defense is missing.

1.2 Re-evaluating “Human in the Loop” in Modern Warfare

In recent years, civil society organizations and AI experts have become increasingly concerned about the emergence of lethal autonomous weapons systems (LAWS), which may use AI to identify targets and thus could harm individuals if used without direct human control (Kahn 2022a). Consequently, there has been a collective initiative at the United Nations to prohibit or impose limitations on using these systems. So far, however, these discussions have not yielded significant results.

Proponents of the “human in the loop” principle, which advocates the ultimate role of humans as decision makers, have noble intentions. But technical solutions envisaged to ensure ultimate human control are challenging to implement in war without overly limiting a defender’s space for action. First, on the battlefield, the distinction between solutions to find, fix, track, and engage targets with a “human in the loop” or in fully autonomous mode is almost impossible to make as the same technology operates in different modes. In practice, there are only 2–5 s between target identification and engagement, which is hardly enough for a human operator to make a balanced decision. This bears the risk that involving a “human in the loop” becomes a “formality,” that might render the target-engagement cycle more difficult. Second, implementing “human in the loop” via mechanisms that ensure remote control of assets not only increases the costs of the respective devices. Russia’s frequent use of electronic warfare (EW) also makes connectivity with the platform and the payload unfeasible. Third, there are no effective international mechanisms to verify or ensure that both sides refrain from using weapons without a “human in the loop,” as it is challenging—if not impossible—to ascertain after a mission has been executed, whether an unmanned asset has been piloted or flown autonomously. Under these conditions, fully autonomous AI solutions would provide the defender with significant advantages in executing its tasks more effectively and more precisely than, for example, with traditional artillery or conventional weapons (Goncharuk 2020).

In sum, the use of AI in the current war is very likely to shape the future regulatory debate in two different ways. First, war experience will test the binding power of norms and rules when aggressors feel unrestricted by the normative principles that guide a defender’s actions. In addition, this will also prompt the need to rethink how the human in the loop logic—which depends on connectivity between operators and unmanned assets—would be ensured under adversarial electromagnetic spectrum dominance.

2 Developing Defense AI

Before the full-scale invasion in 2022, Ukraine’s defense sector was considered non-prestigious and corrupt (Bondar 2023). The defense-industrial complex of Ukraine consisted mainly of state-owned enterprises with only a handful of private defense manufacturers getting state defense orders (Kolomychenko 2023). The main AI development centers were concentrated in a few state institutes that historically worked with the main customer UkrOboronProm and dozens of private companies.

In 2022, most AI initiatives, including data collection and processing, neural network training for weapon detection, and social media analysis to fight disinformation campaigns, were spearheaded by private companies and volunteers. These private and volunteer-led efforts incorporated advanced technologies such as AI-enhanced detectors and trackers in drones and robots for localization, adapting rapidly to the evolving demands of the conflict, whereas the state continued to focus on conventional weapons production. However, as will be discussed below, systemic changes at the state level, including the development of funding programs and the creation of new structural units with a focus on AI development, began in 2023.

2.1 AI’s War Stimulus

Russia pays significant attention to the development of AI technologies in defense. The arsenal of the Russian armed forces already includes a wide range of weapons using AI technologies (CNA 2022). As Katarzyna Zysk (Zysk 2023) shows, Russia’s political and military decision-makers understand the importance of AI and are therefore actively investing in its military application, involving state and private development centers.

Given that Russia is ramping up its industrial production of drones and robotic systems, autonomous robots with AI may begin to be used massively from 2025 to 2026, which could provide a significant advantage on the battlefield. These systems may be EW-proof due to AI’s capability to autonomously process and react to electronic threats, dynamically adapt to changing electronic environments, and operate independently without the need for remote signals, which are often targets of electronic jamming. Consequently, Ukraine is forced to compete and develop AI solutions that provide adequate responses to Russia’s challenges.

From 2023, the Government of Ukraine began actively stimulating the development of innovative defense solutions. The Ministry of Strategic Industries and the Ministry of Digital Transformation are at the forefront of these efforts. Technological defense developments in the field of AI receive organizational and financial support through specialized platforms and projects. In addition, volunteers—who joined to help the Ukrainian armed forces by raising funds for the purchase of weapons, special equipment, transport, communication means, and training—are of pivotal importance in strengthening the country’s defense.

This combination also helps explain Ukraine’s success in implementing AI on the battlefield, which is driven by the motivation of Ukrainian troops, a high level of technical training of military specialists (who are recruited from the private sector), and the involvement of private companies in spearheading novel defense solutions. Crowdsourcing, involving ordinary citizens collect important information about military operations shared with Ukrainian authorities, also plays an important role. Together, these aspects illustrate the broader trend of democratizing military power, which goes far beyond the current war (Kahn 2022b).

Moreover, the unprecedented speed of implementing AI technologies into military systems and their use on the battlefield became possible thanks to the simplification of bureaucratic mechanisms to approve new technologies for use in the defense sector. So far, however, neither Russia nor Ukraine has officially acknowledged deploying fully autonomous weapon systems using AI. This also highlights, that it is “impossible to know, based on open-source materials, whether and what type of AI and autonomous technologies are being used in classified tasks and missions, and to what effect” (Konaev 2023).

2.2 Ukraine’s New Defense AI Priorities

In September 2023, the Expert Advisory Committee on AI Development at the Ministry of Digital Transformation (Expert Committee 2023) provided recommendations for directing state and private investments to accelerate the development of defense AI solutions between 2023 and 2026. Among other things, the committee identified the following defense AI priorities:

  • AI for Domestic Unmanned Systems (e.g., Drones and Robots)

The use of AI is meant to support navigation systems (without Global Positioning System, GPS), coordinated task execution management systems, autonomous task performance systems, enemy weapon and equipment detection (identification) systems, and data collection and storage systems for unmanned systems. These AI applications will be important to augment the “Army of Drones,” that the General Staff of the Armed Forces and the Ministry of Digital Transformation intend to establish to leverage Ukraine’s experience in using unmanned systems to repel Russia’s aggression (Saballa 2023).

  • AI to Combat Disinformation

AI shall be used for interactive solutions based on generative AI, deep voice and deepfake technologies for special law enforcement agencies, automatic detection systems for sources of disinformation and bots involved in foreign influence operations (FIMI), automation/standardization of information threat description and data exchange, labeling and detecting enemy generative AI to prevent the spread of disinformation, and creating datasets and data collection for generative AI development.

  • AI for Defense and Security-Relevant Logistics Systems

AI is used to support predictive maintenance of equipment based on operating conditions, simulation analysis of supply logistics operations, systems for predicting supply and logistics needs and risks, management and automation systems for military warehouses, autonomous robots for military cargo delivery and personnel evacuation during active military operations.

  • AI for Mine and Ammunition Detection and Neutralization

AI should support data collection and accumulation systems from satellites, drones, robots, and other sources (e.g., thermal sensors) for minefield detection, identification of combat sites for demining planning, robotic management systems for demining (urban, field, underwater), and quality control of demining systems.

  • AI for Cybersecurity and Information and Communications Technology (ICT) Protection in the Defense Sector

AI should support systems for radio reconnaissance and advanced EW, modern data encryption and exchange systems, generative AI (voice) for pentesting (e.g., to verify military authorization procedures), countering sophisticated social engineering using generative AI (e.g., voice, 3D video), research focused on developing necessary innovative cybersecurity systems to protect critical digital infrastructures using advanced AI technologies for automatic threat analysis and classification.

  • General Conditions for Rapid Development of AI Solutions for Security and Defense

Accelerating the use of AI for national security and defense foresees simplifying import and licensing procedures for components needed for AI solution training and development, deregulating and simplifying data acquisition procedures from the battlefield for drone, robot, and other AI system developers, simplifying testing processes for innovative products “on the battlefield” by domestic and foreign companies, advancing transparency and accessibility of information for defense tech participants, implementing universally accepted data exchange and storage standards, providing legal-regulatory provision for data exchange between different market participants and state bodies.

  • General Infrastructure and Solutions

Moreover, Ukraine also wants to push for automated combat management systems (automated and AI assistance), simulation modeling solutions for military operations (wargames and military operations research), data analysis and classification systems from surveillance cameras, data collection and accumulation systems from media resources (as part of intelligence data), facial recognition and identification systems, AI-based damage assessment systems with various data types, advanced physical data transmission systems (for drones, robots, between different agents), datasets accessible to market participants for AI training and data collection systems, testing ranges for domestic and foreign developers, data exchange systems between different departments, integration with government business intelligence production of domestic sensors such as stereo cameras, thermal cameras and others.

2.3 Ukraine’s New Defense AI Ecosystem

Many state agencies, private companies, and volunteer initiatives are involved in developing defense AI. As Ukraine’s defense ecosystem grows, it also becomes more complex thus prompting the need for coordination and synchronization. This is an aspect that will need to be strengthened in the future to make sure that ongoing development efforts meet current—and future—warfighting needs.

In addition to the leading Ukrainian ministries, which will be discussed in the next chapter, several para-governmental agencies, a growing number of domestic defense startups as well as international AI companies flocking to Ukraine constitute the country’s defense AI ecosystem. Adjacent to the public sector is the state-influenced defense industrial complex that harbors many of the traditional defense companies. The following entities are of particular relevance:

  • JSC Ukroboronprom

JSC Ukroboronprom is a strategic manufacturer of weapons and military equipment in Ukraine, uniting enterprises in the strategic sectors of the state’s defense industry (Ukroboronprom Undated). Ukroboronprom includes the state enterprise Antonov, which developed more than 100 types of passenger, transport, and special aircraft, like the Ruslan and Mriya platforms (Skorobogatova and Kalko 2016). Ukroboronprom also includes the enterprise State Kyiv Design Bureau Luch, which is one of Ukraine’s leading developers of components for aviation and anti-tank weapon systems (State Kyiv Design Bureau Luch Undated).

  • Brave1 Platform

The Brave1 Platform (Brave1 Undated), established by the leadership of the Ministry of Digital Transformation, acts as an accelerator and an incubator. The main goal of the platform is to assess and promote technological products that can be integrated into various state sectors. In addition, Brave1 offers grants to accelerate the development of technologies considered critically important for Ukraine (Pylypiv 2023). Brave1 supports research projects in the field of AI in priority areas like systems and weapons, protection and security, support and logistics, unmanned systems, cybersecurity, intelligence, and navigation. Based on the military cluster of Brave1, the Griselda system was developed using AI to collect intelligence data. This technology is already in use with other systems like Delta (situational awareness), Bronya (artillery fire support), Kropyva (planning support), and GISArta (C2 for artillery fire) (Denisova 2023).

  • Innovation Development Accelerator

Under the auspices of the Ministry of Defense of Ukraine, the Innovation Development Accelerator (Undated) was created to combine and develop the capabilities of different units inside the Ministry of Defense. This program is designed as a channel for rapid deployment of innovations in the military sphere.

  • Ukrainian Startup Fund

The Ukrainian Startup Fund (Ukrainian Startup Fund Undated) was established to promote the development of early-stage technological startups in Ukraine. Since the beginning of the full-scale invasion, the fund has focused on supporting defense tech and deep tech projects. The fund allocated USD8.2M in grants, funding 352 startup teams, and conducted over 330 events like hackathons, boot camps, educational lectures, or mentoring sessions with 119 startups. The fund has also provided support to attend some of the world’s largest technology events and enabled close to 200 pro bono experts to join the fund’s work (Tarasovsky 2023a, b).

Private funds and non-profit organizations complement the government’s efforts to advance Ukraine’s defense industrial base. The following organizations play a key role:

  • D3 military tech Accelerator

The D3 military tech Accelerator supports investments in early-stage startups, acceleration, and mentorship (Boshnyakov 2023).

  • Flyer One Ventures

Flyer One Ventures (Undated), located in Ukraine, invests in Ukrainian startups from Europe and North America.

  • Aerorozvidka

Aerorozvidka (Undated) is a non-profit organization that promotes the creation and implementation of networked and robotic military capabilities for the Ukrainian security and defense forces.

  • “Come Back Alive” Army Assistance Fund

The “Come Back Alive” Fund, established in 2014, has already raised over UAH 5.2bn (USD 138M) for the needs of the Armed Forces of Ukraine. This fund specializes in the procurement of technical means and was the first in Ukraine to obtain permission to import dual-use goods from abroad (Melezhik 2022).

  • Prytula Foundation

Serhiy Prytula’s charity fund provides aid to the Armed Forces of Ukraine and humanitarian aid to civilians (Serhiy Prytula Charitable Foundation Undated).

  • Ukraine Defense Fund

The Ukraine Defense Fund (Ukraine Defense Fund Undated) provides supplies and supports to people fighting on the front lines of Ukraine.

Ukrainian ministries and programs are meant to strengthen Ukraine’s defense industrial base. Consequently, the number of Ukrainian startups that provide solutions for defense and national emergency management is growing rapidly. Some of these companies also work on defense AI as the following examples illustrate:

  • Artellence, an IT company, has developed neural networks for text analysis, face profile search, and algorithms for cleaning and structuring large data.

  • Athlon Avia, a Ukrainian research and production company, develops unmanned aerial systems for tactical military missions such as intelligence, surveillance, target acquisition, and artillery fire adjustment (A1-CM “Furia”).

  • Saker develops affordable AI for small businesses, including drone-based vision systems for plant protection. The SAKER SCOUT drone, equipped with AI, is used in Ukraine for autonomous target detection and engagement.

  • UA Damage developed a platform for analyzing damage and pyrotechnic contamination using AI, satellites, and drones. Its comprehensive analysis allows visual inspection of unexploded rockets and mines, detecting them in grass and underground, and even localizing explosive objects and marking them on maps.

  • Zvook company creates neural networks of acoustic sensors for detecting aerial objects. The network helps conventional air defense systems detect and destroy enemy cruise missiles and drone attacks and provides alerts to civilians. Zvook technology is trained on a unique dataset of acoustic imprints for most modern Russian weaponry (Ministry of Digital Transformation of Ukraine 2023).

Finally, a unique aspect of military operations in Ukraine lies in the unprecedented support offered by foreign companies. In general, defense collaboration with the United States has been essential for Ukraine. For example, Ukraine has benefited from SpaceX providing access to real-time data transmission in areas where IT infrastructure had been damaged. General Atomics and AeroVironment contributed to advancing surveillance, reconnaissance, and precision strikes. CrowdStrike and FireEye helped advance Ukraine’s digital defense in cyberspace. In addition to financial support, collaboration is meant to advance Ukraine’s domestic defense industrial base and includes agreements on data-sharing and co-production of defense solutions, also with European partners.

Moreover, foreign AI solutions support and enhance Ukraine’s situational awareness and situational understanding regarding Russian threats (Fontes and Kamminga 2023). One notable contributor is Palantir, which played a pivotal role during the conflict by providing extensive data sharing capabilities to the military leadership. Through its MetaConstellation too, Palantir facilitated large-scale data integration, encompassing sources ranging from commercial satellites to classified information from foreign intelligence services. This empowered Ukraine and its allies to utilize commercial data pertaining to specific battlegrounds, enabling the analysis of combat operations and the strategic planning of military endeavors. Notably, during the liberation of Kherson, the Ukrainian military benefited from accurate intelligence on the movements of Russian forces, allowing precise long-range strikes. The information, processed by NATO outside Ukraine, was then transmitted to the Ukrainian command (Ignatius 2022).

American companies such as Planet Labs, BlackSky Technology and Maxar Technologies also contribute significantly by generating satellite imagery of ongoing combat operations and sharing the data with the government and Armed Forces. Furthermore, Primer, an IT company, adapted its voice-to-text program to efficiently process intercepted messages from the Russian military. This innovative approach spares the Ukrainian military from spending extensive hours manually deciphering intercepted conversations. The Primer Command technology facilitates the creation of AI models that swiftly process substantial volumes of data derived from enemy radio communications (Fontes and Kamminga 2023).

3 Organizing Defense AI

Ukraine’s war effort leverages decentralized organizational structures and volunteer initiatives to create a vibrant ecosystem of actors that jointly work on complex systems. This approach also nurtures competition for the best ideas and products between different actors. Although beneficial in principle, decentralized efforts create additional layers of complexity related to coordination and synchronization of activities. In this context, several ministries are involved in setting the course for long-term force planning, operational planning to conduct the war, and strengthening the local defense industrial base. Among the public stakeholders, the following are of particular relevance:

  • Ministry of Defense of Ukraine

In 2021, a Joint Directive of the Minister of Defense of Ukraine and the Commander-in-Chief of the Armed Forces of Ukraine led to the creation of the Center for Innovation and Defense Technologies (Ministry of Defense 2021a). Among other things, the Center has been tasked to develop, test, and provide support for new software that meets the needs of the Armed Forces of Ukraine (Defense Express 2021).

The General Staff of the Ministry of Defense is responsible for military management and develops the country’s defense plans. The General Staff is also in charge of strategic planning of the use of the Armed Forces of Ukraine and certain elements and resources of other parts of the Armed Forces. In addition, the Office for the Development of Automation has been set up within the General Staff. This office has been tasked to organize, coordinate, and execute efforts to advance automated defense solutions, create a unified automated control system, and advance the digitalization of Ukraine’s Armed Forces (Ministry of Defense 2021b).

The Main Intelligence Directorate complements these activities with a focus on conducting intelligence operations relevant for capability development, military construction as well as cybersecurity and military-technical security (Verkhovna Rada 2020).

  • Ministry of Digital Transformation

The Ministry of Digital Transformation implements the state’s digitalization policy and is responsible for policies in robotics and AI. In 2020, the Ministry had established an Expert Advisory Committee on AI Development, whose goal is to enhance Ukraine’s AI competitiveness. This committee, initially the only governmental body with AI expertise, played a significant role before the war. It focused on educational activities, organizing conferences and training sessions to educate military personnel and other agencies about AI applications relevant to their tasks. With the onset of the war, the Expert Advisory Committee shifted its focus to providing consulting support for various tasks. Additionally, some committee members joined the frontline, contributing directly to the defense efforts.

  • Ministry of Strategic Industries

The Ministry of Strategic Industries is intended to manage the military-industrial complex with a focus on producing drones and robots and boosting Ukraine’s indigenous missile program (Press Service of the Office of the Verkhovna Rada of Ukraine 2023).

Ukraine’s organizational setup to advance defense AI evolves commensurate with the threat landscape that shapes and refines the country’s defense AI understanding. As argued above, Ukraine’s defense AI priorities have become more focused reflecting a maturing understanding of the benefits of AI on the battlefield. Nonetheless, as the war goes on, more emphasis will need to be put on how to synchronize the short cycles of battlefield adaptation, that quickly turn instant lessons identified into product upgrades, with the more medium and long-term cycles that look at the necessary changes of concept and capability development that are indispensable to sustain warfighting. This will also have an impact on defense AI governance and the role of coordinating entities that bring the decentralized group of stakeholders together to make sure that everybody understands, supports, and further develops the long-term defense AI priorities discussed above.

4 Funding Defense AI

According to the Stockholm International Peace Research Institute (SIPRI), Ukraine significantly increased its military spending in 2022 by 640% to USD44bn (Tarasovsky 2023a, b). For comparison, Russian military spending grew by 9.2% to USD86.4bn, which is 4.1% of its GDP (Radio Svoboda 2023). Especially as of 2023, Ukraine accelerated funding for AI development. However, due to the war, some program details cannot be disclosed:

  • On 19 October 2023, Ukraine’s Parliament approved the first reading of the 2024 state budget, allocating nearly UAH 56bn (USD1.5bn) for the Ministry of Strategic Industries (Minstrategprom 2023). The 2024 budget also provides UAH2.55bn (USD60M) for the Ministry of Digital Transformation’s digitalization, six times more than in 2023. The approved budget for 2024 includes an increase in funding for defense industry development, specifically for weapon production worth UAH51bn (USD1.3bn) as well as unmanned aerial, surface, and underwater systems worth UAH43.3bn (USD1.1bn) (Ministry of Finance 2023).

  • BRAVE1 launched an AI startup funding program in 2023, awarding 84 defense tech grants worth USD1.53M for AI and other technologies. The platform aims to fund Ukrainian startups with over USD2M by the end of 2023. The 2024 budget foresees a significant increase of the budget to more than USD39M. As of late 2023, an average grant is about USD 5000–25,000, and in 2024, grants will reach amounts of USD50,000, USD100,000, and USD250,000 (TechUkraine 2023).

  • The State Innovative Financial Credit Institution (State Innovative Financial Credit Institution Undated) systematically invests in innovative developments through competitions among universities, institutes, and other state structures. As of 2022, it started including AI projects in dual-use sectors, but specific data is not available.

  • International VC Funds (Stojkovski 2023) are actively monitoring the Ukrainian market, as it serves as a testing ground for creating and testing new weapons, with the goal of investing or engaging on mergers and acquisitions (M&A). Among others, Ukrainian AI startup “Swarmer” secured funding from D3 Accelerator, a fund co-managed by Eric Schmidt. Swarmer’s flagship product AI Copilot enables a single operator to fly a swarm of drones thus boosting battlefield capacities (Yarova 2023).

  • In addition, international companies open local branches in Ukraine to speed up technology development initiatives. German drone manufacturer Quantum Systems, for example, attracted substantial funds with its recent Series B funding round and has decided to join the special economic zone in Diia City to shift parts of its production to Ukraine (Crumley 2023).

Overall, the funding infrastructure for defense innovations focused on AI in Ukraine is in its early stages, actively engaging in processes to establish international cooperation. Considering the high value of local players’ experience, a wave of small and medium-sized M&A of Ukrainian developers by global defense market players is likely to occur within the next years.

5 Fielding and Operating Defense AI

Ukraine’s AI defense ecosystem operates within a rapid development-to-deployment cycle, characterized by the integration of government and military efforts to recruit top engineering talent. This strategy offers engineers an alternative to frontline service, allowing them to contribute to defense projects in secure settings. The approach indicates a strategic use of civilian expertise in military applications, balancing defense development needs with the protection of skilled personnel.

There’s a notable cultural and operational shift as engineers and managers from the civilian IT sector adapt to the traditional military culture. This transition is significant, blending civilian technological expertise with military operational practices, which is essential for developing relevant and practical AI solutions.

A key feature of the ecosystem is the real-time combat testing of AI technologies. Unlike controlled environments, this rigorous testing allows for immediate feedback and rapid iteration, crucial for practical application and improvement of AI in military contexts. It underscores a focus on developing solutions that can withstand the dynamic and unpredictable nature of modern warfare.

However, the ecosystem faces challenges in system integration due to the diversity of systems and standards, affecting the speed and efficacy of AI implementation. This complexity points to a need for standardized protocols and integration strategies to streamline the process.

Many AI solutions, while showing promise, are still in their early developmental stages. This immaturity can lead to increased long-term operational costs, emphasizing the ongoing development, refinement, and investment required to fully realize the potential of these technologies. The ecosystem’s characteristics reflect a strategic blend of civilian and military expertise, rapid adaptation in real-world scenarios, and the ongoing challenge of integrating diverse technologies into a cohesive defense strategy. Key projects and initiatives include:

  • Situational Awareness Solutions

The Delta platform is a situational awareness and battlefield management system (Rosengren 2023). It consolidates information from various sources, including personnel, officials, sensors, and drones. This system is crucial for planning military operations, coordinating unit movements, and securely exchanging enemy location data. Delta was created at the Center for Innovation and Development of Defense Technologies of the Ministry of Defense. The platform is used by intelligence to recognize and identify objects such as enemy soldiers or military equipment. Delta played a pivotal role in destroying the Russian Cruiser “Moskva” in April 2022 (Krasnomovets 2022), as the system enabled the military command to assess the situation in a short period of time and decide to destroy the military target. The Delta system was created according to NATO standards and is compatible with similar situational awareness systems of NATO members (Dobrovolsky 2022).

“Kropyva” is another situational awareness system developed by the volunteers of “Army SOS.” Based on target localization by drones, Kropyva provides fire supports to gunners by computing ballistic calculations. The system receives daily updates on frontline developments. The Ukrainian Army uses Kropyva in combination with other systems like GISArta and more (Melnyk 2022).

  • Data Set Collection Projects

Students at the Ukrainian Catholic University’s Faculty of Applied Sciences (APPS UCU) prepared a specific dataset for object recognition. Data is collected from open sources and can be used by experts to prototype and validate hypothesis (Applied Sciences Faculty of UCU Undated). Kaggle, another database of more than 50,000 public datasets such as equipment losses, for example, can be used for AI training (Kaggle Undated).

  • Disinformation and Bots

With the help of AI, the Mantis analytics platform monitors and analyzes the information space to detect disinformation (Sobachynskyi 2023).

In sum, Ukraine’s AI defense sector is rapidly evolving, marked by a shift towards accelerated development cycles and a focus on integrating advanced AI technologies into defense strategies. Defense AI and other novel defense solutions benefit in particular from

  • easier access to combat data, which manufacturers need to negotiate with the government, to continually refine their products and train AI more effectively.

  • the efforts of the Ministry of Defense to support a community-driven response to streamline the streamline certification processes that significantly reduce the time from an average year to just 1–2 months for drone certification.

  • international companies that establish service and development centers in Ukraine, enabling quick resolution of issues and reduced maintenance time.

  • the fact that key end-users are creating general requirements for defense solutions, providing clear long-term development objectives, and reducing market uncertainty.

6 Training for Defense AI

Amid the ongoing conflict with Russia, Ukraine’s educational landscape has been significantly disrupted. The constant destruction of educational institutions and private residences, coupled with frequent power outages and the tragic loss of family members, have profoundly impacted the quality and accessibility of education. As of June 2023, Russian bombardments have damaged 3290 Ukrainian educational institutions, destroying 262 of them (Ministry of Education and Science 2023). This destruction accounts for about 10% of the country’s educational infrastructure (Center for Economic Strategy 2023).

Despite these challenges, Ukraine trained over 7000 experienced AI developers as of 2022, with its education system producing about 27,000 science, technology, engineering, and mathematics (STEM) graduates each year. Before the full-scale invasion in 2022, AI education in Ukraine typically distinguished between civilian and defense applications. However, the onset of the war prompted a significant shift, with almost all academic programs incorporating military-related components. This shift led to close collaborations between universities and the military sector on various projects.

A noteworthy aspect of Ukraine’s educational response to the war has been the grassroots development of numerous training programs and courses, such as drone operations. These initiatives were often driven by charitable foundations and volunteers, rather than state funding. This approach reflects a unique, community-driven response to the urgent need for specialized training in war time.

Ukraine’s higher education institutions, including the Ukrainian Catholic University, Taras Shevchenko National University of Kyiv, National Technical University of Ukraine “Igor Sikorsky Kyiv Polytechnic Institute”, Kyiv School of Economics, and American University Kyiv, have adapted their programs to meet these new challenges. These adaptations include offering courses and degrees in Data Science, machine learning, and AI that are increasingly tailored to address defense needs.

Given the destruction of the civilian education infrastructure, the military training ecosystem for drone pilots and users of systems like Delta and Kropyva has become even more important. The rapid development of these programs, primarily facilitated by volunteers, underscores the agility of Ukraine’s response to the evolving demands of the war. Military units of the Air Defense Forces, for example, use a Virtual Reality (VR) simulator for military training. This synthetic training environment ensures a more effective preparation of pilots to shoot down enemy cruise missiles, drones, and aircraft (Press Service ODA 2023).

Based on current war experiences, Strata-22 has developed immersive multimedia platforms to train infantrymen, gunners, and tankers. Military operators learn how to handle weapons with the help of largescale models using VR glasses. This simulation environment is built according to NATO standards (Ossipova 2022). In addition, Ukrainian IT company Logics7 has created a universal training system for the fire training of Ukrainian military units. Right now, the system provides digital replicas of different weapon systems like the Next Generation Light Anti-Tank Weapon (NLAW), Javlin anti-tank systems, as well as Stinger and Ingla anti-aircraft systems (Pylypiv 2023). Finally, the war has also highlighted the effectiveness of AI laboratories established at universities, under the aegis of the Ministry of Digital Transformation and the AI Expert Committee. These labs have become instrumental in developing AI tools for wartime use.

In summary, the war has catalyzed a significant transformation in Ukraine’s educational system. Traditional university programs have swiftly pivoted to focus on AI applications in military contexts. Meanwhile, the private sector and non-profit initiatives have stepped in to fill the immediate needs for specialized training and program development, demonstrating remarkable adaptability and resilience in the face of war.

7 Conclusion

War needs require Ukraine to focus on what works best on the battlefield while considering future warfighting requirements as well. This has prompted a clear focus on delivering the best solutions to the frontline and making sure to nurture the talent pipeline for technology development. By contrast, ineffective concepts and burdensome procedures and regulation are pushed aside.

After 2 years of intensive war fighting, it becomes more and more obvious that Ukraine and Russia actively compete in enhancing existing defense solutions with AI and providing new AI-enhanced systems. Against this background, Ukraine is increasingly focusing on (Committee on AI of Ukraine 2023):

  • investing in AI to enhance the autonomy and effectiveness of drones and robotic systems, in particular by enabling these systems to operate without GPS and engage adversarial drone systems;

  • expanding the use of sophisticated sensors, including thermal and real-time satellite video feeds, thereby using for data analysis and data fusion to support decision-making for complex combat operations;

  • developing fully autonomous systems to engage autonomous adversarial systems, thereby using AI to enhance autonomy;

  • establishing new research and development centers to underline the government’s commitment to advancing AI and other defense technologies;

  • integrating AI into all levels of military operations, and

  • making AI solutions more affordable to accelerate a broader and quicker adoption of defense AI by the country’s Armed Forces.

It is important to note that in the ongoing war, Ukraine’s defense AI solutions will be significantly improved thanks to continuous training based on battlefield experience and data. The Ukrainian AI ecosystem has evolved rapidly due to the war, transitioning from grassroots volunteer and commercial efforts to more organized government-backed initiatives and funding. This shift reflects the government’s strategic commitment to leveraging AI for defense. The ecosystem now robustly incorporates R&D, academia, private sector collaboration, and significant government investment, focusing on dual-use technologies for military and civilian benefits. This approach underlines a comprehensive, multi-faceted advancement in AI capabilities driven by the urgency of war.

These advancements are also very likely to shape the future international discussion on the use of defense AI on the battlefield. In this regard, NATO countries should first monitor AI developments in Ukraine and Russia during the war and reflect upon the likely consequences for their own capability planning as well as R&D priorities. Second, it can be assumed that violent non-state actors will also closely follow this war and consider how to best benefit from defense AI. This prompts the need to reconsider how to prevent and limit the proliferation of AI-relevant technologies to non-state actors, which broadens the regulatory debate. Third, countries will also need to be aware that many of the existing data and privacy protection ideas are unfit for warfighting when the comprehensive collection, assessment, fusion, and sharing of public and private data might produce distinct advantages in fighting an adversary. Fourth, this should also lead to a more realistic understanding and reconceptualization of the “human in the loop” principle that will be essential for future certification standards as well defense AI compliance and verification regimes.

Finally, even advanced countries should not underestimate the benefits of crowdsourcing, open-source technology, decentralization, and volunteer efforts that tend to be overlooked in a more mature institutional environment. As these efforts have been crucial in Ukraine’s early and ongoing defense efforts it would be worth analyzing how these efforts can infuse defense innovation elsewhere.