In recent years, the integration of Artificial Intelligence (AI) into military Command and Control (C2) systems and procedures has seen significant progress. This integration is reshaping decision-making processes, enhancing situational awareness, and improving operational efficiency within increasingly complex combat environments. In 2025, the NATO Command and Control Centre of Excellence (C2COE) published the report “AI in Military C2-Systems: An Introduction and Recent Advances.” This report traces the evolution of AI from a theoretical concept to practical military applications, analyzes its role across various stages of the C2 cycle—including data sensing, processing, comprehension, and decision support—and introduces current AI advancements such as Project Maven and AI-driven drone warfare systems. It explores how AI systems can revolutionize capabilities in target identification, intelligence analysis, and overall battlespace awareness, while emphasizing the growing reliance on AI in military decision-making. Additionally, the report discusses key challenges in military AI applications, including data quality issues, cybersecurity vulnerabilities, and the need for explainable AI systems in military contexts. The authors advocate for a balanced approach to AI implementation, leveraging its advantages while retaining human command authority. The fusion of AI with military C2 is described as an evolutionary rather than revolutionary change, necessitating careful consideration of doctrinal adjustments, training requirements, and human-machine teaming models.
1 Introduction
AI is rapidly sweeping across the globe, with major technology companies significantly increasing investments in its practical applications. Recognizing and harnessing the potential of this innovative, breakthrough, and rapidly developing technology within the military domain is of critical importance. This article explores the integration of AI into the military C2 cycle, focusing on its role in improving decision quality, situational awareness, and operational efficiency, while also noting its limitations. The first section examines how AI support systems can be applied across various stages of the C2 cycle, from improved data sensing to enhanced situational understanding. The second section investigates the current state of AI application in the C2 domain, considering advancements in automation, machine learning, and decision support tools, with a focus on its role and impact in recent research and conflicts (such as Project Maven and drone warfare). Finally, the article assesses the broader impact of AI on the C2 cycle, advocating for a balanced and adaptable approach to effectively and responsibly integrate AI.
2 AI Support Systems in the C2 Cycle
NATO defines command and control as the exercise of authority by a commander over assigned and attached forces in the accomplishment of a mission. This term encompasses two elements, command and control, implying a synergistic relationship between them. Command is often associated with human factors like leadership, creativity, and flexibility, while control is more linked to strict rules, doctrine, predictability, and standardization. A primary focus for commanders and their C2 structures is maintaining situational awareness and responding through military operations when necessary to achieve strategic and overarching governmental objectives.
Modern military operations are extremely complex, with data and actions having significant impacts on the battlefield. Military C2 must operate within a rapidly evolving multi-domain environment. To remain effective, militaries must evolve and adapt alongside technological development. In particular, AI as a transformative force can enhance decision quality, operational efficiency, and strategic capability, but its integration into C2 systems requires careful consideration. Vigilance is needed against technological solutionism—the belief that technology alone can solve all challenges without considering the complex social, cultural, and political factors affecting the military and security environment. A RAND report highlighted that commanders must understand the limitations of AI in five key areas: cybersecurity, predictive maintenance, wargaming, mission planning, and constructive simulation. Addressing these five areas is crucial for strengthening AI as a reliable and effective tool in military operations.
AI has the potential to improve all levels of the overall C2 framework and decision-making process, though this improvement has limitations regarding decision authority. Research from the NATO C2COE indicates that while the technology for fully autonomous (primarily at the tactical level) military systems exists, the commander’s role remains essential, as they wish to make the final decisions. Trust in support systems—whether human staff or AI algorithms at different levels—is key to a commander’s work. This also influences the delegation of authority and task assignment within the C2 process. Work is underway in related areas, such as DARPA’s Explainable AI (XAI) program, which aims to enable commanders to understand, trust, and effectively use machine learning technologies. Deeply understanding decision models is vital for effectively integrating AI into command processes.
(1) Decision Models – Components of the C2 Cycle
A widely used model is Boyd’s Observe-Orient-Decide-Act (OODA) loop. The OODA loop provides a comprehensive overview of the fundamental premise of effective C2: an entity (whether an individual or organization) that can process this cycle faster than an opponent—observing and reacting to unfolding events more quickly—can “get inside” the opponent’s decision cycle and gain an advantage. The OODA loop shows that before making a decision (the Decidestage), a commander must first collect information (Observe) and determine its meaning and possible actions (Orient). “Shortening the loop” can become decisive.
The C2 cycle (or NATO C2 Conceptual Model), developed as part of the C2 Capstone Concept, organizes the elements of C2 differently from the OODA loop and presents the associated steps more clearly. By decomposing Boyd’s broad and multi-faceted concept, the C2 cycle allows for a more detailed analysis of its components and a clearer articulation of how emerging technologies like AI can enhance decision quality within this framework. Although the steps of the C2 cycle are depicted as equally important, in reality they are not and do not require the same time investment. As with the OODA loop, the speed of completing the entire C2 cycle is crucial for gaining operational advantage.
The core of the C2 cycle model (from inner to outer circle) is “Connect,” which is the enabling element of C2, linking and coordinating the three phases of the C2 model: Sense, Decide, and Act. In this context, the capabilities provided by AI align with and can enhance all aspects of the C2 cycle.
Based on core design functions, AI systems can be applied as decision support tools in three main functional areas: 1) Description & Analysis; 2) Prediction & Inference; and 3) Recommendation. Consequently, the first and second functional areas are primarily related to the data sensing and processing stages, while the second and third are more closely linked to the comprehension and decision stages of the C2 cycle. Although these areas are not strictly separate and may overlap, they collectively cover the core functions of AI systems.
(2) Improved Data Sensing and Processing
Just as air superiority is crucial for conducting full-spectrum air operations, information superiority (or dominance in the information domain) is equally important. The application of AI plays a key role in achieving this information advantage and in processing vast amounts of data rapidly and accurately. Adopting these technologies enables the faster acquisition and dissemination of information among allies and partners.
In the past, military commanders had to rely on staff working laboriously on manual information processing, as well as their own intuition and experience to make decisions. These decisions were mostly based on relatively limited data. The increasingly widespread use of emerging technologies for sensing and data collection leads to data overload, which is difficult for humans to process fully. While gaining more insight is generally beneficial, too much information can lead to decision paralysis. AI can help address this by filtering, correlating, and fusing data. The application of advanced algorithms and machine learning techniques allows AI systems to uncover connections and correlations that analysts might miss. By identifying patterns and detecting anomalies using current events and historical data, AI will alleviate the cognitive burden on decision-makers. However, this also highlights the potential vulnerability of human analysis, making it a weak link in the decision loop.
Despite its many advantages, AI still faces significant challenges requiring further research. One key challenge in fully integrating AI into C2 systems is handling unstructured data. Additionally, mission uncertainty, assessing opponent intent through heuristic intervention, small sample sizes, data inconsistency, high-clutter environments, heterogeneous inputs, opponent manipulation in adversarial and deceptive environments, explainability, and meaningful human control remain critical hurdles.
To enable AI to effectively handle massive data volumes, it is necessary to guide AI in autonomous learning and processing of classified data; extracting relevant data and transforming it into usable information and intelligence, prioritized appropriately. A promising approach to meeting these needs is the use of Large Language Models (LLMs). These models will continue to transform the military and defense sectors by improving decision-making (e.g., providing creative course of action suggestions), enhancing situational awareness, and overall operational efficiency. They can integrate data from various sources such as text, images, and sensors, providing comprehensive insights by analyzing satellite imagery, interpreting intelligence reports, and monitoring social media for threats in real-time. This holistic approach will help defense personnel make informed decisions quickly, shortening response times and improving mission outcomes. Furthermore, LLMs will improve communication and coordination among teams, ensuring accurate transmission of critical information. Integrating multimodal agents into defense operations will significantly elevate the role of AI in national security.
To ensure effective decision-making processes, data needs to be distilled to its core elements as it moves up the chain of command, without omitting relevant facts. Therefore, Information and Knowledge Management (IKM) is an integral part of C2, aimed at enhancing situational awareness. Contextual interpretation of data involves integrating analyzed information into the broader framework of the mission, objectives, and operational environment. This step is crucial to prevent misinterpretation of raw data, ensuring factors like geography, weather, local customs, and culture are considered in information analysis. Collected data must be processed and analyzed to extract meaningful insights, identify patterns, trends, and potential threats, and assess the credibility and relevance of information. These processes are iterative and adaptable, meaning analysis may need revision and updating as new data emerges. High-quality information processing and analysis contribute to enhanced situational awareness, information superiority, and decision quality.
(3) Enhanced Situational Understanding
A comprehensive understanding of the operational environment does not necessarily mean having the most sensors or the largest datasets. True cognitive advantage comes from the ability to understand data and project it within specific contexts and mission frameworks, forming situational understanding. To achieve this, NATO doctrine recommends applying the Comprehensive Understanding of the Operational Environment (CUOE) process. This is a cross-headquarters process for a crisis, aimed at developing a comprehensive understanding covering all Political, Military, Economic, Social, Infrastructure, and Information (PMESII) domains, including relevant potential threats, risks, and opportunities, to support operational planning and execution as part of a broader campaign.
Given the enhancements in data sensing and processing, AI enables further analysis that will ultimately contribute to better environmental understanding. AI can assist specialized staff in accelerating the provision of guidance, thereby helping commanders make well-informed decisions in a timely manner. These systems possess a robust knowledge base (data lakes) built on historical encounters and collaborative databases. The recommendations or judgments they provide based on this knowledge base help build trust and understanding. Commanders need to be able to significantly shorten the time of the decision-making process while understanding the effectiveness limits of specific systems. They can then take steps to minimize the impact of potential errors. Military leaders bear responsibility for their decisions, and this responsibility does not change even if AI systems support those decisions.
The application of AI will significantly impact decision-makers’ ability to process and integrate a wide variety of information sources. However, whether a top-down or bottom-up approach is used, it is humans who train these AI/ML systems, and ultimately, humans should be accountable. From a NATO perspective, it is recognized that the core of future military advantage lies in effectively integrating humans and machines into operational teams. This will create a fusion situation: staff and/or AI-based systems assess information in specific areas of interest, providing actionable insights for situational understanding. Shared situational awareness in Human-Machine Teaming (HMT) requires dialogue between humans and computers through intuitive interfaces. To cope with the vast amount of data in military conflict, human mental capacity is insufficient, while computer algorithms face challenges from uncertainty and ambiguity in data and decisions. Methods to improve this process involve equipping specialized anomaly detection tools and correlating events across domains. Non-human intelligent collaborative systems with “human-on-the-loop” capabilities are needed to provide situational understanding, operational assessment, and alternative analysis, supporting commanders in achieving comprehensive understanding of the operational environment. Developing and managing Information Requirements Management and Collection Management (IRM & CM) systems, including the Commander’s Critical Information Requirements (CCIR), is crucial for improving the Military Decision-Making Process (MDMP). The increasing complexity of data highlights the future importance of commanders’ and staff’s skill in “knowing what to ask,” in other words, the correct information needed for operations. On this basis, the following section will illustrate current applications of AI in C2.
3 Current AI Developments in the C2 Domain
At the operational level, AI systems play a significant role in streamlining processes related to targeting, including detection, validation, nomination, and prioritization. Advanced AI-driven platforms, such as the U.S. Project Maven, Palantir’s MetaConstellation software, Ukraine’s “Griselda” system, Russia’s “Bylina” electronic warfare system, Ukraine’s “Kropyva” system, Russia’s “Acacia-M” system, Israel’s “Alchemist” system, Ukraine’s GIS Arta system, Israel’s “Gospel” system, “Lavender” system, and “Where’s Daddy” system, integrate real-time intelligence, sensor data, and pattern analysis to enhance situational awareness and optimize military decision-making. This distributed situational awareness, created through collaboration between human and technological teams, contributes to increased efficiency and effectiveness. These systems can identify and track potential targets, assess their relevance based on operational criteria, and prioritize engagement based on threat level and mission objectives. By processing massive amounts of data from multiple sources like satellite imagery, geolocation tracking, and drone reconnaissance, decision-makers utilize AI to identify targets, process data and intelligence, and assess the legitimacy of potential targets. Reports on current conflicts (such as the Israel-Hamas war, the Russia-Ukraine war, and the Syrian conflict) highlight the increasingly widespread use of AI in military decision-making. Furthermore, the U.S. Department of Defense’s Project Maven demonstrates the integration of AI-driven analytics in intelligence operations, enhancing target identification and battlespace awareness, further solidifying the dependence on AI in military decision-making.
(1) Project Maven
At the tactical and operational levels of the C2 cycle, AI supports complex decision-making in force employment by integrating state-of-the-art computer vision and AI capabilities into analytical workflows. These workflows include tasks such as geolocating targets, identifying anomalous activities in near-real-time, detecting anomalies, and recognizing targets. AI application at this stage is crucial because manually analyzing millions of visual records is impractical.
Historically, the process of targeting in warfare—whether sensor-to-shooter or inform-to-effects—developed slowly. Identifying and locating targets, tracking them, and passing this information to weapon systems took considerable time, especially during the Cold War. By the 1990s and early 2000s, processing reconnaissance data often took 15-20 minutes, with additional time required to deploy fire platforms. By today’s standards, this process can be completed in just a few minutes under optimal conditions. However, earlier methods required large-scale targeting centers and numerous personnel. For instance, during Operation Iraqi Freedom in 2003, a single targeting center employed over 2,000 staff.
In response to the scale and speed demands of modern conflict, the U.S. Department of Defense initiated Project Maven in 2017 to shorten response times and enable synchronized, parallel execution. By 2022, the U.S. National Geospatial-Intelligence Agency (NGA) took over the project, transforming it into an operational military capability. The project aims to use computer vision and AI algorithms to identify targets in real-time from existing data (like drone footage) and social (online) dynamics. Maven processes incoming data, applies AI to detect Points of Interest (POIs), and generates battlefield overlays showing potential targets, friendly forces, and civilian infrastructure. The final engagement decision is then made by warfighters. As part of this project, the U.S. military began testing an AI-driven targeting system for reconnaissance data analysis in 2020. Unlike traditional methods, this system compares newly captured images with existing databases to identify targets even under imperfect data conditions. While operators remain indispensable, AI increases the speed and precision of targeting decisions.
Project Maven functions similarly to a large-scale facial recognition system, processing video feeds to identify and track targets, distinguishing real threats from decoys. Project Maven has been trained on millions of reference points across different reconnaissance scenarios, enabling it to make predictions and calculate probabilities. Today, it is integrated into battlefield command systems, processing complementary data streams from commercial satellite imagery, military sensor networks, and intelligence databases to suggest targeting strategies. In 2020, the project demonstrated its capability by identifying and locating a target in under a minute, a task that previously took 12 hours using traditional methods. Since the outbreak of the Russia-Ukraine war in 2022, Project Maven has played a significant role in processing visually collected satellite information, which has been shared with Ukrainian forces. Furthermore, this conflict provided a valuable testing ground for the system, with Project Maven undergoing 50 rounds of improvements during the war.
Despite its success, Project Maven’s AI faces challenges, with further refinements planned. In 2023, U.S. Central Command’s Chief Technology Officer, Schuyler Moore, acknowledged that the project’s AI performed poorly in determining optimal attack sequences and selecting the most appropriate weapons. While AI excels at target identification, decision-making remains a challenge due to the need for creativity and human judgment. Therefore, officials continue to insist that AI will not be granted the authority to autonomously make fire decisions. Throughout the C2 cycle, there are currently no plans to delegate the Decidestage to such systems. Therefore, it’s important to note that these systems remain part of the data sensing and processing stages and do not extend beyond the decision stage, while ensuring that all critical decisions ultimately remain under human control and follow current ethical, moral, and legal frameworks.
(2) Drone Warfare
In recent years, drones have garnered significant attention, especially through real-time video streams shared via social media. Conflicts like the Second Nagorno-Karabakh War and the Russia-Ukraine war have highlighted the impact of drones on the battlefield. In contrast, while the integration of AI systems into military operations deepens, public records remain relatively scarce. Most information comes from reports, news articles, and official statements. Some emerging media have reported in detail on the role of AI in the Russia-Ukraine conflict, providing insights into the development and deployment of AI-driven First-Person View (FPV) drones. These developments indicate that drone warfare is increasingly shifting towards autonomy and AI assistance.
AI-driven FPV drone warfare is becoming a game-changer, not only in the Russia-Ukraine conflict but also in ongoing operations in northern Syria. AI-driven drone swarms, along with cutting-edge cyber and Electronic Warfare (EW) capabilities, have become central pillars of modern military operations. As these technologies develop, so do the measures needed to counter them. For example, to address the dual threat of cyber and drone attacks, Ukraine is testing the “Stowaway” counter-drone system. This system was developed by two leading U.S. tech companies: IronNet, an AI-based cybersecurity company, and Asterion Systems, which focuses on counter-drone system technology capable of detecting drones, classifying them, tracking their trajectory, jamming counter-drone system networks, and destroying target drones. In 2024, Ukraine deployed approximately 1.5 million drones incorporating domestic AI systems. AI is primarily used to enable drones to autonomously reach targets without direct piloting, maintaining effectiveness even in areas with heavy electronic interference. While achieving 100% target identification accuracy is unlikely, success rates of 80% to 90% could still have a significant impact in large-scale conflicts, especially when striking thousands of targets simultaneously. Combined with mass-produced low-cost missiles and FPV drones (like those increasingly used in the Russia-Ukraine war), AI-driven targeting systems have the potential to transform modern warfare. Both the U.S. and China are leveraging industry to significantly expand their inventories of missiles and FPV drones.
However, the role of AI in military operations is not limited to these early stages; it extends to adaptive decision-making, operational planning, and the execution phase. As AI continues to evolve, its integration throughout the entire C2 cycle—including predictive modeling, automated response suggestions, and real-time mission adjustments—will redefine the future of C2. While human oversight remains crucial, AI’s expanding capabilities are transforming military decision-making, making operations faster, more precise, and increasingly autonomous. Current AI applications, such as Project Maven and drones in conflict zones, primarily focus on data analysis, processing, and comprehension in the initial stages of the C2 cycle. These AI systems are trained to identify targets, process vast amounts of sensor data, and assist in decision-making. However, as shown in the C2 cycle framework in Figure 1, AI’s influence extends beyond the Comprehendstage into the Decide, Act, and Assessstages, reinforcing information superiority, execution superiority, and overall operational effectiveness.
4 Conclusion
AI-driven decision-making is not merely a technological advancement; it requires changes to command structures, doctrine, and processes. Furthermore, it demands that operators acquire new skill sets. Only by addressing these issues can AI-enhanced C2 realize its potential to transform military operations.
Currently, AI primarily serves as a tool for enhancing the capabilities of armed forces, provided its application is optimized for efficiency, effectiveness, and agility. However, the challenges of military innovation persist in the AI domain. Future research can focus on addressing current challenges in military AI applications.
AI enhances decision-making by accelerating the C2 cycle, improving intelligence analysis, operational planning, and data processing. To fully realize its potential, armed forces must integrate AI efficiently and effectively, ensuring it serves as a force multiplier rather than a replacement for human judgment. This requires a shift in mindset from skepticism or partial adoption to active adaptation, modifying traditional approaches to incorporate AI-driven insights. For successful implementation, military personnel must cultivate data and technology literacy, trust in digital systems, algorithmic reasoning skills, and AI-assisted decision-making skills, ensuring that AI-generated outputs are correctly interpreted and integrated into command processes.
To enable future staff to adopt decision support products, platforms like drones and Project Maven, and (near-)real-time simulations, the provided solutions should be available on-demand and have a low threshold for practical use. To fully leverage technology-enabled decision support systems, achieving true interoperability of processes and procedures is essential. Developing effective ways of working and utilizing available tools will help build confidence in integrated technologies.
As AI technology advances, armed forces must remain adaptable and willing to refine their approaches. Armed forces can leverage AI to revolutionize their operations and maintain strategic advantage.