This dissertation investigates the impact of Artificial Intelligence (AI) on contemporary military conflicts at the tactical-operational and strategic levels of war. While the use of AI systems has rapidly proliferated in the military domain, it remains underexplored within the disciplines of International Relations and Security Studies from an empirical perspective. This thesis aims to address this gap while also amending the tendency of existing literature to overstate AI’s revolutionary potential and overlook its technical limitations. To do so, it adopts a qualitative methodological framework grounded in the offense-defense theory to provide a structured lens for evaluating how AI systems impact military operations. This research is carried out by means of an empirical case study analysis focusing on three cases: the deployment of Project Maven AI systems by the United States in their operations against international terrorism, Ukrainian armed forces’ use of military AI in the Russo-Ukrainian War, and the use of AI systems by Israeli Defense Forces in their 2021 and 2023 military operations against Hamas in Gaza. This analysis uses observable implications to assess the pervasiveness of AI systems within decision-making processes, and their impact on operational tempo and mission success rate; while accounting for intervening variables such as enemy countermeasures. The findings show that AI systems shift the balance towards offense-dominance at the tactical-operational level by accelerating decision-making, improving target acquisition, and shortening the sensor-to-shooter cycle. However, this impact does not spill over to the strategic level, where the influence of AI remains limited. These findings challenge the dominant assumptions about AI’s revolutionary impact on warfare and reframe the debate by grounding it in observable operational effects, rather than solely in AI’s speculative potential. The nuanced understanding of AI’s impact herein presented is critical for scholars, policymakers and military professionals alike, as it also offers insights applicable to the study of other emerging and disruptive technologies. This dissertation’s contribution is both theoretical and practical: on the one hand, it offers an empirically grounded analysis of AI’s military applications, underscoring the need for further research to assess the evolving role of AI and its military, political, and ethical implications. On the other hand, it stresses the need for careful evaluation of the ways in which AI can be integrated into military decision-making processes and, consequently, the need for human oversight over such systems.
Questa tesi indaga l’impatto dell’intelligenza artificiale (IA) sui conflitti militari contemporanei a livello tattico-operativo e strategico. Sebbene l’uso dei sistemi di IA si sia rapidamente diffuso in ambito militare, questo rimane poco esplorato all’interno delle discipline delle Relazioni Internazionali e dei Security Studies da una prospettiva empirica. Questa tesi mira a colmare questa lacuna e a emendare la tendenza della letteratura esistente a sopravvalutare il potenziale rivoluzionario dell’IA e a trascurare i suoi limiti tecnici. Per farlo, l’elaborato di ricerca adotta una metodologia qualitativa basata sulla teoria dell’attacco-difesa, fornendo un quadro strutturato per valutare l’impatto dei sistemi di IA sulle operazioni militari. La ricerca è condotta attraverso un’analisi empirica di casi di studio incentrata su tre casi: l’impiego dei sistemi di IA di Project Maven da parte degli Stati Uniti nelle operazioni contro il terrorismo internazionale, l’uso dell’IA militare da parte delle forze armate ucraine nella guerra russo-ucraina e l’uso dei sistemi di IA da parte delle Israeli Defense Forces nelle operazioni militari del 2021 e 2023 contro Hamas a Gaza. L’analisi utilizza implicazioni osservabili per valutare la pervasività dei sistemi di IA all'interno dei processi decisionali e il loro impatto sulla velocità operativa e sul tasso di successo delle missioni, tenendo conto anche della presenza di variabili intervenienti come le contromisure nemiche. I risultati mostrano che i sistemi di IA spostano rendono l’equilibrio più prono alla preponderanza dell’attacco a livello tattico-operativo per mezzo dell’accelerazione de processi decisionali e del miglioramento l’acquisizione dei bersagli. Tuttavia, tale impatto non ingenera un pari effetto a livello strategico, in cui l’influenza dell’IA rimane molto limitata. Questi risultati mettono in discussione le ipotesi dominanti sull’impatto rivoluzionario dell’IA sulla guerra e riformulano il dibattito basandolo su effetti operativi osservabili, invece del solo potenziale speculativo dell’IA. La comprensione dell’impatto dell’IA qui presentata è d’interesse per accademici, policymaker e militari professionisti, poiché offre riflessioni applicabili anche allo studio di altre tecnologie emergenti. Il contributo di questa tesi è sia teorico che pratico: da un lato, offre un’analisi empiricamente fondata delle applicazioni militari dell’IA, sottolineando la necessità di ulteriori ricerche per valutarne il ruolo in evoluzione e le implicazioni militari, politiche ed etiche. Dall’altro, sottolinea la necessità di un’attenta valutazione dei modi in cui l’IA può essere integrata nei processi decisionali militari e, di conseguenza, la necessità di supervisione umana su tali sistemi.
ARTIFICIAL INTELLIGENCE AT WAR: THE IMPACT OF MILITARY AI SYSTEMS ON THE OFFENSE-DEFENSE BALANCE
Russo, Alessandra
2025
Abstract
This dissertation investigates the impact of Artificial Intelligence (AI) on contemporary military conflicts at the tactical-operational and strategic levels of war. While the use of AI systems has rapidly proliferated in the military domain, it remains underexplored within the disciplines of International Relations and Security Studies from an empirical perspective. This thesis aims to address this gap while also amending the tendency of existing literature to overstate AI’s revolutionary potential and overlook its technical limitations. To do so, it adopts a qualitative methodological framework grounded in the offense-defense theory to provide a structured lens for evaluating how AI systems impact military operations. This research is carried out by means of an empirical case study analysis focusing on three cases: the deployment of Project Maven AI systems by the United States in their operations against international terrorism, Ukrainian armed forces’ use of military AI in the Russo-Ukrainian War, and the use of AI systems by Israeli Defense Forces in their 2021 and 2023 military operations against Hamas in Gaza. This analysis uses observable implications to assess the pervasiveness of AI systems within decision-making processes, and their impact on operational tempo and mission success rate; while accounting for intervening variables such as enemy countermeasures. The findings show that AI systems shift the balance towards offense-dominance at the tactical-operational level by accelerating decision-making, improving target acquisition, and shortening the sensor-to-shooter cycle. However, this impact does not spill over to the strategic level, where the influence of AI remains limited. These findings challenge the dominant assumptions about AI’s revolutionary impact on warfare and reframe the debate by grounding it in observable operational effects, rather than solely in AI’s speculative potential. The nuanced understanding of AI’s impact herein presented is critical for scholars, policymakers and military professionals alike, as it also offers insights applicable to the study of other emerging and disruptive technologies. This dissertation’s contribution is both theoretical and practical: on the one hand, it offers an empirically grounded analysis of AI’s military applications, underscoring the need for further research to assess the evolving role of AI and its military, political, and ethical implications. On the other hand, it stresses the need for careful evaluation of the ways in which AI can be integrated into military decision-making processes and, consequently, the need for human oversight over such systems.File | Dimensione | Formato | |
---|---|---|---|
tesiphd_completa_russo.pdf
accesso solo da BNCF e BNCR
Dimensione
1.48 MB
Formato
Adobe PDF
|
1.48 MB | Adobe PDF |
I documenti in UNITESI sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
https://hdl.handle.net/20.500.14242/210731
URN:NBN:IT:UNICATT-210731