AI, Cybersecurity, and the Future of Military Drone Software

By Isaac Liu and Kat Patten

Abstract: 

As the internet of things (IoT) continues to grow and rapid advancement of artificial intelligence (AI) transforms both civilian and military environments, systems become increasingly digitized, interconnected, and data driven. Artificial intelligence (AI) revolutionizes modern military operations by driving the evolution of software-defined defense on unmanned aerial systems (UAS). Military drones become more dependent on AI-powered accelerators, communication networks, and onboard computing. This reshapes how defense systems function, communicate, and respond to evolving threats, contributing to national intelligence and security defense. While these innovations and transformations enhance operational capabilities, they have also expanded the landscape of cybersecurity threats, increasing the demand to mitigate cyberthreats. This blog post investigates the interconnected relationship between AI, cybersecurity, and military drones by exploring both the opportunities and challenges created by AI integration. It discusses how AI strengthens military cybersecurity through intelligent threat detection, autonomous decision support, edge computing, and software assurance while simultaneously introducing new cybersecurity risks, including adversarial attacks, data poisoning, GPS spoofing, and sensor manipulation. The blog also examines the ethical implications of incorporating AI into autonomous military systems, emphasizing the importance of maintaining human oversight, algorithmic transparency, and compliance with international humanitarian law to minimize civilian harm. Ultimately, this blog argues that AI should serve as a force multiplier rather than a replacement for human decision-makers. As software-defined defense continues to reshape modern warfare, the successful deployment of AI-enabled military drone systems will depend on balancing technological innovation with robust cybersecurity protections, ethical governance, and continuous human oversight to maintain national security and the strategic advantage of U.S. intelligence community. 

INTRODUCTION

In today’s fast-paced, digital, and data-driven environment, artificial intelligence (AI) has become deeply integrated across civilian sectors. AI has transformed healthcare through medical image analysis, education through adaptive learning and personalized instruction, and the business domain through predictive analytics, and AI-assisted software development. The military has embraced AI as part of its ongoing digital transformation. Military applications of AI range from decision support systems to national security operations such as counterpiracy, counterterrorism, and border security. The future of warfare is no longer defined solely by tanks, fighter jets, and soldiers at the frontline. As military operations become increasingly autonomous and digitally interconnected, AI enables faster threat detection, improves decision-making, increases operational efficiency, strengthens intelligence analysis, supports autonomous operations, powers predictive simulations, and accelerates incident response.[1] This convergence of software and defense has given rise to software-defined defense, a new paradigm that integrates artificial intelligence, cybersecurity, and digital technologies into the deployment of modern weapon systems. 

Among the technologies enabled by software-defined defense, unmanned aerial systems (UAS), commonly referred to as drones, have emerged as one of the most transformative innovations in modern warfare. UAS play a critical role in maintaining U.S. deterrence and preserving its global military leadership. Their extensive use during the Russia-Ukraine War demonstrated how drones can conduct persistent surveillance and precision strikes while minimizing risks to military personnel.[2] Ukraine’s deployment of these weapons throughout the conflict illustrates the critical role drones play in modern warfare.[3] Drone operations have enabled Ukraine to continuously conduct high-precision strikes and position the battlefield in their favor.[4] Increasingly, these platforms rely on AI to support autonomous navigation, target recognition, mission planning, and real-time decision-making.[5] Collectively, these outcomes underscore the strategic importance and value of continued investment in AI-enabled drones and secure autonomous systems. 

The U.S. military has increasingly incorporated software-intensive technologies into modern weapons systems, particularly unmanned aerial systems.[6] AI is being used to improve the reliability, autonomy, and effectiveness of these systems. However, their reliance on software, communication networks, and data connectivity introduces new cybersecurity vulnerabilities. At the same time, the same technologies that enable greater autonomy also expands the attack surface for cyber threats.[7] Each additional software component creates another potential attack vector, exposing drone systems to threats such as adversarial attacks, data poisoning, model theft, GPS spoofing, sensor manipulation, and malicious software updates.[8] Consequently, securing AI-driven drone software requires protecting not only traditional computing infrastructure but also the integrity of the data, algorithms, and decision-making processes that underpin autonomous operations.[9] As these platforms become increasingly autonomous and software dependent, protecting them against cyber threats has become equally important as improving their capabilities. This has opened the door to exploring how AI can be incorporated into developing cybersecurity solutions for military software. This blog examines how AI-enabled drone software simultaneously enhances military capabilities while introducing new cybersecurity risks, arguing that AI-driven cybersecurity solutions, ethical governance, and adaptive regulatory frameworks are essential for the secure and effective deployment of autonomous military systems. 

SECTION I: THE DUAL ROLE OF AI IN MILITARY DRONE SYSTEMS 

While AI-enabled drone systems have transformed military operations by improving autonomy, intelligence, and operational effectiveness, these same technologies have fundamentally changed how cybersecurity must be approached. Understanding both the defensive capabilities AI provides and the new vulnerabilities it introduces is essential for securing modern software-defined defense systems.

Modern unmanned aerial systems are no longer simply remotely piloted aircraft; they are highly interconnected cyber-physical systems that depend on software, communication networks, sensors, and artificial intelligence to perform mission-critical tasks.[10] Their ability to navigate with minimal human intervention; identify targets and strikes on massive battlefield data converted from images and videos; and exchange information faster and in real time greatly enhances military effectiveness.[11] However, this reliance on software and communication networks also makes cybersecurity a fundamental requirement for ensuring mission success and protecting national security.[12]

A significant advancement in modern UAVs is the transition from physical systems and cloud-assisted computing to onboard embedded artificial intelligence through edge computing and powerful AI accelerators/processors.[13] Earlier generations of unmanned systems relied heavily on transmitting sensor data to ground stations or cloud-based platforms for image analysis, target recognition, and decision support.[14] Although effective in permissive environments, this architecture proved vulnerable to communication latency, bandwidth limitations, and electronic warfare tactics such as jamming or signal disruption. With edge computing (i.e., data processing framework that transforms computation and processes data at or near where the data is generated originally rather than delivering to far cloud servers in traditional cloud computing for computing) and AI accelerators (i.e., tiny device that runs large language models (LLMs), deep learning, and reinforcement learning models locally without cloud connectivity, servers, or high-end GPUs) computationally intensive tasks like computer vision, object detection, terrain classification, target tracking and strikes, and autonomous route planning happens locally onboard in real-time.[15] Those powerful processors foster the drone’s ability to learn through trial and error iteratively and “thinking” autonomously. To reduce dependence on continuous communication with external systems and cloud connectivity, embedded AI enables unmanned aerial systems to operate and continue executing mission objectives even in contested or denied electromagnetic environments.[16] This architectural shift provides decisive cybersecurity and operational advantages. By computing locally, it minimizes the reliance on vulnerable communication links that opponents can intercept, jam, manipulate, or spoof.[17] Despite disconnections or signal loss, drones equipped with onboard AI accelerators continues to compute and execute pre-authorized actions, such as “avoiding obstacles, tracking targets, returning to base, or completing reconnaissance tasks.”[18] By these means, this transforms, matures, and contributes/augments UAVs as an essential asset and force multiplier to the military as well as software-defined defense rather than a merely blind and useless platform once a drone lose connection. 

At the same time, however, concentrating intelligence within the drone itself expands the attack surfaces.[19] As drones “absorb”/evolve with more capabilities and expanding into a supercomputer “hive mind”, each “edge” is an opportunity to attack. AI models, code, onboard processors, firmware, and all other technologies become high-value targets for adversaries to manipulate and corrupt training data and eventually disarm and/or seize control of the aircraft. Consequently, “cyberattacks against drones results in loss or manipulation of sensitive data, unauthorized surveillance, hijacking of flight control systems, and even the repurposing of drones for malicious operations.” [20] Hence, securing the onboard AI stack is vital and as demanded by safeguarding the communication network that integrates military platforms on drones. 

To combat that risk, Project Mayhem, an AI-based autonomous bot checker tool, was created by Carnegie Mellon University’s startup ForAllSecure after winning Pentagon’s research agency DARPA’s Cyber Grand Challenge back in 2016.[21] The success and works of Project Mayhem directly translate into modern AI-enabled unmanned aerial systems because military drones are fundamentally software-defined platforms on lines of code and structural logic. Every function and integration depends on running and stable millions of lines of code. A single software flaw makes the system dysfunctional, exposes the attack opportunity 

The project transforms cyber defense from manual, time-intensive hacking to a more efficient and scalable machine-driven security system. In the past, programmers, software bug hunters, security coders, and other cybersecurity professionals took hours to review vast amount lines of codes to search, locate, and fix malfunctions and vulnerabilities.[22] This is very impractical. Now within minutes they can crack down on security risks.[23] The decision-making, validation, and process still require human expertise input; however, it amplifies security professionals to get more work done as computing power tremendously increased and more efficient time are saved and redirected. It provides application and cloud API/Code/SBOM security testing solutions that identify defects and vulnerabilities in software code, further enhancing the protection of data.[24] The software generates thousands of tests per minute and has consistent reinforcement of learning and self-learning on its past runs and interactions.[25] It acts as a reflective agent that improves software security and mitigates risks just like a security professional, if not even more. 

In fact, Mayhem could locate software system risks and vulnerabilities in almost all weapon systems within the military between 2012 and 2017 Department of Defense reported in 2018.[26] Mayhem “hacks” and detects several forms and techniques: 1) fuzz testing, 2) symbolic execution, 3) automated triage and reproduction, and 4) regression testing. The fuzzing security testing technique involves Mayhem inundating the “target software with randomly generated input (e.g., commands or photos), resulting in memory leaks, crashes, and software vulnerabilities.”[27] The symbolic execution involves analyzing program through Mayhem probing all possible execution paths, running multiple test cases, and representing inputs and variables as symbolic mathematical expressions to detect potential weak spots.[28] Mayhem automatically recreates the identified vulnerabilities in a controlled experiment, allowing development teams to inspect and fix issues to remove false positives (i.e., condition is present when it’s not).[29] For regression testing, Mayhem runs multiple executed test cases up against corrected codes to validate and verify fixes that are stable and maintained, preventing reoccurrence of bugs again after adjustments.[30]   

Long before a drone is deployed into combat, these testing capabilities strengthen and check for cybersecurity against unexpectedness and malicious attack scenarios/simulations.[31] This proactive approach allows defense organizations to address and mitigate software weaknesses during development and maintenance phases, thereby increasing the resilience, reliability, and mission-capability of UAVs. 

SECTION II: THE ETHICS OF INCORPORTING AI INTO MILLITARY DRONES & CYBERSECURITY CONFLICT

Although AI offers significant military and cybersecurity advantages, there remains considerable debate regarding the ethical implications of integrating AI-powered software into military technologies such as drones. Oftentimes, developers aren’t aware of the impact their designs have on an ethical level. At a webinar hosted at the University of Waterloo’s AI institute, it argued that one of the hardest global governance challenges to tackle was incorporating AI into war and conflict.[32] An important area of ethics in the military domain is prioritizing human life.[33] While AI-enabled drones can reduce risks to military personnel, they may also increase the likelihood of unintended civilian causalities if autonomous systems make incorrect targeting decisions. This ethical analysis is referred to as Jus War theory in international politics. Jus War Theory is the ethical standards to follow for going to war and while in war.[34] More specifically, jus in bello lays out the fundamental ethics of what behaviors are and are not ethical during war.[35] This included how weapons of war such as AI powered drones are used.  

While AI itself is not a new concept, the concept of incorporating AI into defense technologies such as military drones is still under review. Although the use of AI can be ethical when it comes to saving human life, it can also raise ethical risk when it is placed in the wrong hands or when it completely compromises human decision-making.[36]Keeping jus war theory in mind, it’s important the humans remain at the center of the decision-making process rather than AI to prevent ethical challenges and irreversible consequences in warfare. If AI is trained on incomplete or biased data, a drone may incorrectly classify a school and civilian for a military facility and target, respectively, increasing the risk of unlawful engagement.[37] It is also important that the commander understands the full scope of the data that the AI is making its decision off.[38] Although AI systems can process battlefield information far more quickly than humans, speed should not come at the expense of accuracy, accountability, or ethical judgment (e.g., civilian causalities and destruction).   There is very little policy and governance surrounding the usage of AI in the context of war which raises ethical concerns as nation state actors and non-state actors have freedom over usage.

One real world application of AI drones in war is the Russia-Ukraine conflict. The conflict demonstrates how the use of technology is what classifies it as ethical rather than the technology itself. Ukraine’s use of drone technology shows that AI-powered recognition of targets on military drones can be used in an ethical way as the targets were oil facilities[39]. Even when AI is incorporated into technology, the use remains ethical because civilian lives are not at risk. However, on the other hand. Russia launched a strike using 351 drones which resulted in 22 civilians being killed.[40]Russia consistently uses its drone technology in civilian areas or areas near civilians. This makes it so that incorporating AI into technology makes it more likely that the system may mistake a military target for a civilian. Thus, making the use unethical.

In addition to the Russia-Ukraine conflict, AI and drones are generally used in the military context for surveillance. RAS sensors can be used to detect injured soldiers on the battlefield. Multimodal-AI can also be used in terms of medicine to monitor the state of a soldier’s injury and keeping records of their vital signs to inform medics.[41]  This allows medics to make the appropriate decision for when it is important to enter the battlefield to rescue the soldier. Importantly, the AI does not make the decision for the medic but rather aids in the medic decision-making process.[42]Instead, it analyzes battlefield conditions and provides decision support. When used in this way, the use of AI in drones is ethical because it helps preserve as many human lives as possible; both the soldiers and the medic have a higher survival rate. This illustrates that the ethical value of AI depends not on the technology itself but on how humans choose to design, validate, supervise, and deploy AI-enabled drone systems. As the human medic is relying on information from the drone, it is important that the algorithm’s decision-making process and analysis is fully transparent and explainable to the human decision-maker.[43] Incorporating AI into drone technology to help save human life positively supports jus un bello (how the weapon is being used in war) as it prevents the loss of human life and the human still has autonomy over the decision. 

Because AI-enabled drones rely on software, sensors, and communication networks, cybersecurity introduces an additional layer of ethical responsibility. The lack of policy to regulate AI leads to room for increased cybersecurity risk and ethical concerns in the cyber domain. [44] AI becomes an ethical challenge when it is used to conduct cyber operations such as spoofing drone sensors, disrupting communication systems, and spreading disinformation during armed conflict. In fact, Russia used in the conflict against war to corrupt Ukraine data.  Russia has consistently launched spoofing cyber-attacks against Ukraine drones which have resulted in the drones being redirected and targeted at Romanian civilian facilities except for their initial target.[45] AI, cybersecurity and ethics become tricky when it comes to when something crosses the line of being “ethical”. In the case of war, AI and cybersecurity become unethical when its civilians are harmed.

To address the ethical concerns surrounding the integration of AI, cybersecurity, and AI-enabled military drone software, the European Union and intergovernmental organizations should continue to propose policy that outlines when states should and should not use AI in military weapons. In addition to taking precautions, states should abide by the Article 36 review.[46] This means that states continue to regularly test their systems as software innovates and evolves.[47] The ethical challenges associated with AI-enabled drones reflect a broader shift toward software-defined defense: as software assumes greater responsibility for mission-critical tasks, ensuring human oversight, transparency, accountability, adherence to international humanitarian law, and cybersecurity become an ethical as well as technical requirement.

From this perspective, AI is considered ethically justifiable when it functions as a decision-support tool that protects human life rather than replacing human judgment in lethal decision-making.[48]

CONCLUSION

Artificial intelligence has fundamentally reshaped modern military operations by transforming unmanned aerial systems into intelligent, software-defined defense platforms through enhancing threat detection, increasing computing power locally, and strengthening the resilience of drone software as well as the technology itself under contested environments within battlespace. AI’s capabilities can enhance U.S. deterrence in strategic advantage and surpass major adversaries in the international system. For the US to remain in a hegemonic power and have military superiority, they need to keep up with cybersecurity capabilities. Integrating AI into military cybersecurity will allow the U.S to continue to surpass major adversaries. Thus, the U.S military should continue investing in incorporating AI into major cybersecurity areas such as analyzing network traffic and conducting projects similar to Project Mayhem through investments, opportunities, funding, programs, and many more channels to gather the research, talents, and development of security interests. 

However, technological progression alone is not enough: the operationalization of AI requires comprehensive and adaptive transitions and strategies that integrate long-term vision, innovation, and ethical responsibility. This is currently and will continue to be an iterative and specifically monitored process that checks AI in its deployment, capabilities, understanding/learning, and development. Human insight is still critical when using AI to solve complex cybersecurity problems. AI should be treated as a force multiplier rather than a replacement for human decision-makers. With both human insight and AI integration, the U.S can maximize their cybersecurity capabilities and solve problems at a faster rate while keeping decision precision. The future of software-defined defense will depend not only on developing more intelligent and autonomous systems but also on ensuring those systems remain secure, ethical, transparent, and resilient against the evolving cyber threats that emerge. With continuous research, funding, and development of AI in military defense and domains, a near goal of the U.S. Department of War could be reached, perhaps in the next five years: JADO (Joint All Domain Operations) and Mosaic Warfare, in which AI is assisting and integrated all five warfighting domains (i.e., air, land, maritime, cyberspace, and space).  

Notes:

[1] Sauser, Maj. Mark K. “Unmanned Aircraftand the Revolution inOperational Warfare.” Military 

Review, 2025.

[2] Maj. Mark K. “Unmanned Aircraftand the Revolution inOperational Warfare.”, 2025.

[3] Maj. Mark K. “Unmanned Aircraftand the Revolution inOperational Warfare.”, 2025.

[4] Maj. Mark K. “Unmanned Aircraftand the Revolution inOperational Warfare.”, 2025.

[5] Soare, Dr Simona R., Dr Pavneet Singh, and Dr Meia N. “Software-Defined Defence:       

Algorithms at War.” The International Institute for Strategic Studies, 2023.

[6] Scully, Tom, Mark Robson, Mohanad Sarhan, Nour Moustafa, and Javaan Chahl. “Software Defined Wide Area Networking for Secure and Resilient Unmanned Aerial Systems.”           Computers & Industrial Engineering 213 (2026)

[7] Scully, Tom, Mark Robson, Mahanad Sarhan, Nour Moustafa, and Javaan Chahl. “Software Defined Wide Area Networking for Secure and Resilient Unmanned Aerial Systems.” Computers & Industrial Engineering Volume 213 (2026). https://doi.org/10.1016/j.cie.2025.111741.

[8] Karpinski and Relich, AI Resiliency Against Adversarial Spoofing and Perturbations.

[9] Kacem, Thabet, and Kensley Benjamin. “Artificial Intelligence Methods for Unmanned Aerial 

Vehicles Cybersecurity: A Comprehensive Survey.” (2026)

[10] Alsadie, Deafallah, Cybersecurity and Artificial Intelligence in Unmanned Aerial Vehicles: Emerging Challenges and Advanced Countermeasures, IET Information Security, 2025, 2046868, 50 pages, 2025. https://doi.org/10.1049/ise2/2046868

[11] Alex Vakulov, “Embedded AI in Military Drones Is Redefining Autonomy and Operations,” IDGA, 2026, https://www.idga.org/government-defense-it-communications/articles/embedded-ai-in-military-drones-is-redefining-autonomy-and-operations.

[12] Alsadie, Deafallah, Cybersecurity and Artificial Intelligence in Unmanned Aerial Vehicles: Emerging Challenges and Advanced Countermeasures, IET Information Security, 2025, 2046868, 50 pages, 2025. https://doi.org/10.1049/ise2/2046868

[13] Vakulov, Alex. “Embedded AI in Military Drones Is Redefining Autonomy and Operations.” IDGA, 2026. https://www.idga.org/government-defense-it-communications/articles/embedded-ai-in-military-drones-is-redefining-autonomy-and-operations.

[14] Vakulov, Alex. “Embedded AI in Military Drones Is Redefining Autonomy and Operations.” IDGA, 2026. https://www.idga.org/government-defense-it-communications/articles/embedded-ai-in-military-drones-is-redefining-autonomy-and-operations.

[15] Vakulov, Alex. “Embedded AI in Military Drones Is Redefining Autonomy and Operations.” IDGA, 2026. https://www.idga.org/government-defense-it-communications/articles/embedded-ai-in-military-drones-is-redefining-autonomy-and-operations.

[16] Vakulov, Alex. “Embedded AI in Military Drones Is Redefining Autonomy and Operations.” IDGA, 2026. https://www.idga.org/government-defense-it-communications/articles/embedded-ai-in-military-drones-is-redefining-autonomy-and-operations.

[17] Vakulov, Alex. “Embedded AI in Military Drones Is Redefining Autonomy and Operations.” IDGA, 2026. https://www.idga.org/government-defense-it-communications/articles/embedded-ai-in-military-drones-is-redefining-autonomy-and-operations.

[18] Vakulov, “Embedded AI in Military Drones Is Redefining Autonomy and Operations.”

[19] Alsadie, Deafallah. Cybersecurity and Artificial Intelligence in Unmanned Aerial Vehicles: Emerging Challenges and Advanced Countermeasures. IET Information Security, vol. 2025, no. 1 (2025): 50. https://doi.org/10.1049/ise2/2046868.

[20] Vakulov, “Embedded AI in Military Drones Is Redefining Autonomy and Operations.”

[21] DARPA (Defense Advanced Research Projects Agency) (United States). “‘Mayhem’ Declared Preliminary Winner of Historic Cyber Grand Challenge.” 2016. https://www.darpa.mil/news/2016/mayhem-winner-cyber-grand-challenge.

[22] S, Tom. “This Bot Hunts Software Bugs for the Pentagon.” WIRED, 2020. https://www.wired.com/story/bot-hunts-software-bugs-pentagon/.

[23] S, Tom. “This Bot Hunts Software Bugs for the Pentagon.” WIRED, 2020. https://www.wired.com/story/bot-hunts-software-bugs-pentagon/.

[24] Mayhem Security. “Mayhem Security: Automated Code and API Security Testing.” https://www.mayhem.security/why-mayhem-security.

[25] S, Tom. “This Bot Hunts Software Bugs for the Pentagon.” WIRED, 2020. https://www.wired.com/story/bot-hunts-software-bugs-pentagon/.

[26] DARPA (Defense Advanced Research Projects Agency) (United States). “‘Mayhem’ Declared Preliminary Winner of Historic Cyber Grand Challenge.” 2016. https://www.darpa.mil/news/2016/mayhem-winner-cyber-grand-challenge.

[27] Mayhem Security. “Mayhem Security: Automated Code and API Security Testing.” https://www.mayhem.security/why-mayhem-security.

[28] Mayhem Security. “Mayhem Security: Automated Code and API Security Testing.” https://www.mayhem.security/why-mayhem-security.

[29] Mayhem Security. “Mayhem Security: Automated Code and API Security Testing.” https://www.mayhem.security/why-mayhem-security.

[30] Mayhem Security. “Mayhem Security: Automated Code and API Security Testing.” https://www.mayhem.security/why-mayhem-security.

[31] Alsadie, Deafallah. Cybersecurity and Artificial Intelligence in Unmanned Aerial Vehicles: Emerging Challenges and Advanced Countermeasures. IET Information Security, vol. 2025, no. 1 (2025): 50. https://doi.org/10.1049/ise2/2046868.

[32] Bessma, Aaron Shull, and Jean-François Bélanger. Introduction_ The Ethics of Automated 

Warfare and AI – Centre for International Governance Innovation.Pdf. 2022.

[33] Lee, P., Ahmad, T., Waheed, S. M. & Kenning, A. An AI ethics framework for a trustworthy autonomous drone system to support battlefield casualty triage. AI Ethics 6, 139 (2026)

[34] Just War _ Carnegie Council on Ethics in International Affairs.Pdf.” Carnegie Council for Ethics and International Fairs, n.d.

[35] Just War _ Carnegie Council on Ethics in International Affairs.Pdf.” Carnegie Council for Ethics and International Fairs, n.d.

[36] Bessma, Aaron Shull, and Jean-François Bélanger. Introduction_ The Ethics of Automated 

Warfare and AI – Centre for International Governance Innovation.Pdf. 2022.

[37] KHACHATRYAN, DAVIT.  Military AI Challenges Human Accountability – CIP.pdf.. Military AI Challenges Human Accountability – CIP.pdf.

[38] KHACHATRYAN, DAVIT.  Military AI Challenges Human Accountability – CIP.pdf.. Military AI Challenges Human Accountability – CIP.pdf.

[39] Reporter, Guardian Staff. “Ukraine war briefing: Drones strike Russia oil refinery in Siberia as Zelenskyy warns region now ‘within reach.’” The Guardian, July 7, 2026. https://www.theguardian.com/world/2026/jul/07/ukraine-war-briefing-drones-strike-russia-oil-refinery-in-siberia-as-zelenskyy-warns-region-now-within-reach.

[40] Russia’s Missile and Drone Attacks on Ukraine Kill at Least 22 in the Kyiv Region _ PBS News.Pdf. N.d.

[41] Lee et al. An AI ethics framework for a trustworthy autonomous drone system to support battlefield casualty triage, 2026.

[42] Lee et al. An AI ethics framework for a trustworthy autonomous drone system to support battlefield casualty triage, 2026.

[43] Lee et al. An AI ethics framework for a trustworthy autonomous drone system to support battlefield casualty triage, 2026.

[44] Pauwels, Eleonore. Civilian Data in Cyberconflict_ Legal and Geostrategic Considerations, 2022.         

[45] Livingstone, Katie. “How Russia Is Turning Ukraine’s Drones Against NATO.” Defense News, June 1, 2026. https://www.defensenews.com/global/europe/2026/05/29/how-russia-is-turning-ukraines-drones-against-nato/.

[46] KHACHATRYAN, DAVIT.  Military AI Challenges Human Accountability – CIP.pdf.. Military AI Challenges Human Accountability – CIP.pdf.

[47] KHACHATRYAN, DAVIT.  Military AI Challenges Human Accountability – CIP.pdf.. Military AI Challenges Human Accountability – CIP.pdf.

[48] Lee et al. An AI ethics framework for a trustworthy autonomous drone system to support battlefield casualty triage, 2026.