The Lawful Use of Autonomous Weapon Systems for Targeted Strikes (Part 2): Targeting Law & Practice

Abstract

Lethal Autonomous Weapon Systems (LAWS) are essentially weapon systems that, once activated, can select and engage targets without further human intervention. While these are neither currently fielded nor officially part of any nation’s defence strategy, there is ample evidence that many States and defence contractors are currently developing LAWS for future deployment. The main IHL problem posed by these weapon systems is that lethal action against specific targets may be taken by sensory hardware and control software, rather than human operators exercising deliberative reasoning and judgment. However, in the case of targeted strikes, commanders and their battle staffs do select specific targets during the targeting cycle, thereby restricting machine discretion to the choice of munition and the timing of weapons release. This ameliorates a significant concern often put forward by ban proponents, which is the lack of meaningful human control in weapons autonomy. In Part 1 of this sub-series of three articles, the concepts, advantages and technologies of LAWS were explained in order to lay down a basic understanding of how and why autonomous weapons (as opposed to remotely-piloted systems) will be used in targeted strikes. After recapping some of these points, this Part 2 article will build on these foundations in three broad sections. Firstly, the Joint Targeting Cycles are explained and applied to situations where the deployment of LAWS may be an option. As will be seen here, when commanders and their staffs select specific targets and decide on precautions in attack, they follow very detailed and deliberative steps. Consequently, within the overall targeting process, machine discretion is largely bounded in favour of deliberative human control. Second, there is a consideration of the ‘meaningful human control’ (MHC) concept and how this is potentially achieved, both ‘downstream’ (in the targeting process), and ‘upstream’ (in design and development processes). Here, it is argued that a) organised and structured decision-making processes during the targeting cycles ensure there is MHC in the ‘wider loop’, b) this acts as a check on ‘narrow loop’ autonomy, and c) due to the ‘individual attack’ limitation, MHC is relatively more easily achieved in a targeted strike. Arguably, these preclude any prohibition on autonomous targeted strikes within an armed conflict and inside the boundaries of a ‘hot battlefield’. Finally, the legal basis for specifically permitting targeted strikes in an armed conflict will be laid down, and applied to the use of LAWS for such operations. This part will invoke legal transplants from the Convention on Cluster Munitions, to argue for both legal permissiveness of autonomous targeted strikes as well as some limiting principles, all in light of the nature and scope of MHC established above. Moreover, there is a consideration of whether and how this legal basis, which is presumed to apply to object and personality strikes, should extend to signature strikes. This sub-series of three articles is the fourth of a seven-part series, which comprises: 1) assessing the sense and scope of autonomy; 2) whether and how LAWS can be designed/deployed in compliance with IHL; 3) issues relating to the explosive remnants of war; 4) the lawful use of LAWS for targeted strikes (current sub-series of three article); 5) legal and policy issues in the autonomous delivery of nuclear weapons; 6) arms control issues; and 7) non-proliferation.