Abstract
Following the September 11, 2001 attacks, the Bush administration's authorization of Unmanned Aerial Vehicles as lethal weapons fundamentally altered the ethical calculus of armed conflict, directing substantial federal funding into autonomous weapons development while outpacing the governance frameworks designed to regulate it. This thesis addresses the gap between those two trajectories in two components that are designed to sit in tension with one another.
The technical component presents the design, development, and evaluation of ICARUS-1, a dragonfly-inspired quad-wing unmanned aerial system. Conventional UAS architectures present well-documented trade-offs: fixed-wing platforms demand continuous forward motion and open operating space, while multirotor systems offer precision hovering at the cost of aerodynamic inefficiency and limited endurance, restricting both configurations in the constrained environments where adaptability is operationally decisive. Dragonflies resolve these trade-offs biologically through independently actuated wings that enable hovering, rearward flight, and rapid directional change. ICARUS-1 translates these mechanics into an engineered system through fully independent four-wing actuation, grounded in Computational Fluid Dynamics modeling and three-axis bench testing. The system functions as a proof of concept that this architecture is achievable, with direct implications for the future design and deployment of military aerial platforms.
The STS component argues that the central justification for autonomous drone warfare, the surgical precision of UAVs, is a political construct rather than an operational reality. Drawing on the methodological framework established in Walzer's Just and Unjust Wars and applied to non-state armed conflict, this research conducts a historical Just War analysis of CIA drone operations in Pakistan and Afghanistan through quantitative open-source strike data. That data reveals civilian casualty ranges wide enough to constitute structural failure under Just War Theory's core pillars of distinction and proportionality. The deeper argument is that Just War Theory was never architecturally equipped to evaluate autonomous systems – its pillars presuppose human judgment at the point of lethal decision, a condition that the delegation of wartime authority to algorithms and remote operators has systematically eliminated. In response, this thesis proposes Just Drone War Theory, a four-pillar framework introducing jus ante bellum dronicum, a pre-deployment pillar requiring algorithmic audits, defined authorization criteria, and mandatory human oversight before autonomous systems are cleared for operational use, enforced through the proposed Autonomous Warfare Authorization Board, a permanent congressional body designed to restore democratic accountability to a domain that has operated under sustained executive opacity since 2001.
What both components make clear is that technical advancement and ethical obligation are not sequential – they are concurrent, and treating them otherwise has produced a responsibility gap that has grown proportionally with the capabilities of the systems themselves. Drone warfare has not made conflict more precise; it has made it more accessible, and the engineers developing that capability are not outside the scope of that consequence. This thesis argues that enforceable standards must precede deployment, and that the professional obligations of engineers do not end at the boundary of technical performance.
