ABSTRACT
In our contemporary moment, computation has unsettled international conflict through an inherent, and growing, incommensurability expedited by ‘artificial intelligence’ (AI) systems. These offer increasingly recursive machine learning capacities that use ‘big’ data to interpret and perform international conflict in distinctly new ways. Computation, this chapter proposes, not only transforms the speed of conflict but also subtly contorts and twists our norms through its alternative capacities for (re)cognition. To explore this, ‘deep’ reinforcement learning (RL) algorithms, popularised by DeepMind’s AlphaGo, are examined and applied to military wargaming and offensive cyber operations to speculate on the often-imperceptible transformation of social norms by calculative means. By embracing the work of French philosopher Emmanuel Levinas on the ethical condition of the face, the chapter argues that deep RL consists of various computational and social traces that articulate a new ethico-political incommensurability – or unknowability – of conflict through, and by, AI. The chapter then complicates controllable and ethically accountable AI systems through the human in, on, and out of the loop by disavowing and unsettling AI as simply a tool as it (re)cognises, performs, and transforms where, who, and how international conflict is enacted.
