Killing With Drones, Proportionality, and Trolley Problems

Reports of killing by drones are increasing. Initially they were exceptional, now they are commonplace. Every few weeks there is a report of another killing, invariably by the US, in some far off country. With the rapid pace of technological development, the investment being made into more and more autonomous systems, and little sign of this trend being checked, it can only continue. The ethicality and legality of such practices outside international armed conflict is extremely dubious. In the context of international armed conflict the practice is checked only by the concept of ‘proportionality’, a concept that is problematic generally, and is almost entirely unable to discharge the heavy burden that is imposed on it by the practice of drone killing.

The concept of proportionality in warfare, both moral and legal, states that when contemplating an attack, the number of expected civilian casualties or damage to civilian objects must be weighed against the military advantage anticipated. A disproportionate weighing is, in principle, both immoral and a war crime. How this weighing is to be undertaken is not transparent. Rather than specifying a clear ratio of what proportion of military to civilian casualties might be acceptable (say 1:1), the weighing process is almost entirely intuitive. This is probably unavoidable: the potentially huge number of different factors that need to be taken into account in making such decisions would be way too complicated to represent with any workable formula.

However, the fact that proportionality is assessed intuitively is not an insurmountable problem. In some other contexts, the principle works tolerably well. For example, the almost all the rights of the European Convention on Human Rights rely on some form of proportionality assessment. That this is possible relies on the fact that that humans – provided they are in a equivalent positions – are remarkably consistent in their judgements.

This consistency can be illustrated by results from ‘Trolley’ Problems. Originally a philosophical thought experiment devised by Foot (1967), Thompson (1986) and others, turning them into actual experiments has become a burgeoning area of research. These originally focussed on the dilemma of a runaway trolley that could be diverted in different ways, resulting in different numbers of casualties, and now feature all manner of hypothetical scenarios. Their advantage over real life moral dilemmas is that the parameters can be very carefully controlled. What is remarkable is the substantial consistency in judgements that is revealed when experimental participants are presented with the same dilemmas.

The problem with proportionality being an intuitive assessment is not therefore insurmountable. The real problem appears to be that those assessing proportionality are not in comparable positions. In other words, though the decision whether to launch a military attack will generally depend on a huge number of different factors, most people would reach a similar decision, provided they were truly considering those same factors. The problem comes where the relevant factors differ. Michael Waltzer, in his famous book, Just and Unjust Wars, gives the example of a warring party facing an existential threat. In such circumstances, the possibility would arise that that party would consider an attack with a huge number of civilian casualties as proportionate. So proportionality becomes whatever that party considers proportionate at the time, and as such is extremely subjective.

But even this problem of subjectivity could be overcome. One strategy would be to have the assessment of proportionality scrutinised by a third party in a less subjective position. This is essentially the role of domestic courts, supervised by the European Court of Human Rights in the context of proportionality decisions under the European Convention on Human Rights. It works tolerably well. Supervising proportionality decisions in warfare is what the International Criminal Court is supposed to do. Yet there is little immediate prospect of the US, or any other very powerful state, submitting to its jurisdiction.

So I also have a more modest proposal of a method to reduce the civilian casualties of proportionality assessments in drone warfare that that may be more feasible. This is suggested by another pattern disclosed by the use of trolley experiments. This pattern is that apparently arbitrary changes to the environment in which the proportionality assessment is made can make significant changes to how much harm is considered acceptable. For example, research suggests that individuals are prepared to tolerate greater harm according to the way the dilemmas are presented. For example, subjects will accept greater harm if the killing results from impersonal, mechanically applied force, rather than personal force applied through immediate bodily contact. While such result would predict a rather depressing outcome of current trends in drone warfare, it may be possible to counterbalance this to some extent. Research is beginning to isolate the environmental factors that result in fewer civilian casualties in proportionality decisions. International conventions could be agreed on the environments in which proportionality assessments had to be made. This might require, for example, that decisions are only made ‘in theatre’, that those making assessments are publically identified (perhaps by a pseudonym), or that those making proportionality assessments also spend a set portion of their deployment on front line combat duties.