Controlling AWS: A Cyclical Process

The ability to control weapons is an essential component of targeting. For autonomous weapon systems (AWS), the notion of meaningful human control (MHC) has frequently been advanced as a framing device for policy and legal discussions, but the concept is functionally void due to lack of precision as to what it entails. This chapter presents the results of an empiric study on the different ways MHC has been conceptualised in literature and discourse. Focal points of influence throughout the targeting process are identified during which commanders can exercise significant influence over their AWS, in the form of Awareness, Weapon Selection, Context Controls, Prediction, and Accountability. In addition, forward processes are identified between these points that suggest a forward progression: for example, Accountability is only possible if the user knows of the risks (Awareness) and makes an informed choice to deploy an AWS in particular way (Weapon Selection and Context Controls). These findings are distilled into an MHC Framework, which outlines the practical steps a commander must take in order to maintain control over their AWS when ordering (iterative) attacks, even if some use-of-force elements are delegated to the AI. The theory is validated by superimposing the MHC Framework’s elements over the NATO Targeting Cycle, which shows that the focal elements correspond well with the steps in the cycle, demonstrating the MHC Framework’s viability as a workable guideline that can be implemented during real military operations.