Probabilities towards death: bugsplat, algorithmic assassinations, and ethical due care

Abstract

ABSTRACT This article explores the principle of due care in war and the myth that improved battlefield technology makes Western warfare inherently more ethical. The discursive construction – which I term virtuous chaoplexic militarism – of the US as ethical by virtue of its utilization of technologically advanced modes of killing, seeks to dissolve the ethico-political dilemmas of war into quantifiable problems to-be-solved. This article illustrates this dissolution by outlining the transformation within US military decision-making from an ethics of practical judgement to a computational techno-ethics. To do this, I evaluate two concrete cases of US algorithms of militarism. The first case traces the rise of collateral damage estimation algorithms, colloquially known as bugsplat. I examine how bugsplat is programmed, its fundamental design flaws, and its practical exploitation by commanders to erroneously tick the box of ethical due care. The second case explores the SKYNET machine-learning algorithm that was designed to construct ‘legitimate targets’ for US drone strikes via heterogeneous correlations of SIM card metadata. While drone strikes are widely praised for their capacity to individualize targeting, the algorithmic process of SKYNET ultimately erodes the individual subjectivity that is foundational for ethics of war through data constructions of ‘terroristness.’ As both cases demonstrate, the ultimate goal of this virtuous chaoplexic militarism is to render the ethico-political dilemmas of killing quantifiable, predictable, and solvable. There exists an urgent need to interrogate socio-technical interactions in the military setting; and specifically, the degree to which practical judgement has been outsourced to a morally problematic computational techno-ethics.