Abstract:
In addition to its successful mobilization in stigmatization and norm-setting processes on anti-personnel landmines and cluster munitions, the principle of distinction as enshrined in International Humanitarian Law also figures prominently in the debate on lethal autonomous weapons systems (LAWS). Proponents of a ban on LAWS frame these as indiscriminate, that is, unable to distinguish between civilians and combatants, and thus as inherently unlawful. The flip side of this particular legal argument is, however, that LAWS become acceptable when considered capable of distinguishing between combatants and civilians. We thus argue, first, that this particular legal basis for the call for a ban on LAWS might be rendered obsolete by technological progress increasing discriminatory weapon capabilities. Second, we argue that the argument is normatively troubling as it suggests that, as long as civilians remain unharmed, attacking combatants with LAWS is acceptable. Consequently, we find that the legal principle of distinction is not the overall strongest argument to mobilize when trying to stigmatize and ban LAWS. A more fundamental, ethical argument within the debate about LAWS – and one less susceptible to ‘technological fixes’ – should be emphasized instead, namely that life and death decisions on the battlefield should always and in principle be made by humans only. Lethal autonomous weapons systems: a threat to human dignity Numerous arguments motivate the current call for an international, legally binding ban on so-called lethal autonomous weapons systems (LAWS). Strategic concerns include proliferation, arms races and escalation risks (Altmann and Sauer, 2017; Rickli, 2018). Military concerns include the incompatibility of LAWS with a traditional chain of command or the potential for operational failures cascading at machine speed (Bode and Huelss, 2018; Scharre, 2016). Ethical concerns include the fear that LAWS might further increase the dehumanization and abstractness of war (and thus its propensity), as well as its cruelty if warfare is delegated to machines incapable of empathy or of navigating in dilemmatic situations (Krishnan, 2009; Sauer and Sch€ ornig, 2012; Sparrow, 2015; Sparrow et al., 2019; Wagner, 2014). Legal concerns include difficulties of attribution, accountability gaps, and limits to the fulfillment of obligatory precautionary measures (Brehm, 2017; Chengeta, 2017; Docherty, 2015). But the most prominent concern, focalizing some elements of the concerns just mentioned, is the danger these weapons pose to civilians. This argument’s legal underpinning is the principle of distinction – undoubtedly one of the central principles of International Humanitarian Law (IHL), if not the central principle (Dill, 2015). As multifaceted and complex as the debate on military applications of autonomy is now, what has been articulated at its very beginning (Altmann and Gubrud, 2004; Sharkey, 2007) and consistently since then is that LAWS would violate IHL due to their inability to distinguish between combatants and civilians. This image of LAWS as a threat to civilians is echoed routinely and placed first by all major ban supporters (we substantiate this claim in the following section). That LAWS would be incapable of making this crucial distinction – and thus have to be considered indiscriminate – is assumed because ‘civilian-ness’ is an underdefined, complex and heavily context-dependent concept that is not translatable into software (regardless of whether the software is based on rules or on machine learning). Recognizing and applying this concept on the battlefield not only requires value-based judgments but also a degree of situational awareness as well as an understanding of social context that current and foreseeable computing technology does not possess. We unequivocally share this view as well as these concerns. And yet, in this article, we propose to de-emphasize the indiscriminateness frame in favor of a deeper ethical assertion, namely that the use of LAWS would infringe on human dignity. The minimum requirement for upholding human dignity, even in conflicts, is that life and death decisions on the battlefield should always and in principle be made by humans (Asaro, 2012; Gubrud, 2012). Not the risk of (potential) civilian harm, but rather retaining meaningful human control to preserve human dignity should be at the core of the message against LAWS. © 2019 University of Durham and John Wiley & Sons, Ltd. Global Policy (2019) 10:3 doi: 10.1111/1758-5899.12691 Global Policy Volume 10 . Issue 3 . September 2019 370 SpeclSection A ricle Our proposal rests on normative considerations and strategic communication choices. In the remainder of this article, we elaborate on two basic lines of our argument, namely the IHL principle of distinction and the concept of human dignity, provide insights into how and why they have been mobilized in the global debate on LAWS, and discuss the benefits and challenges of putting our proposal into practice. LAWS and the principle of distinction Modern IHL identifies three different categories of persons: combatants, non-combatants, and civilians. Those members of the armed forces who directly participate in hostilities count as combatants; those members who do not directly participate (e.g. military clergy) count as non-combatants; and persons who do not belong to the armed forces count as civilians (Aldrich, 2000; Ipsen, 2008). These distinctions bring into being one major principle for the conduct of hostilities: Only members of the armed forces constitute legitimate targets, whereas civilians must never be deliberately made a target of attack (Best, 1991). With regard to the use of certain means and methods of combat, the prohibition of indiscriminate attacks implies a prohibition of indiscriminate weapons. Weapons may be deemed indiscriminate if they cannot be targeted at specific and discrete military objects, if they produce effects which cannot be confined to military objects during or after the use of the weapon, or if they are typically not targeted at specific objects despite being capable of precise targeting in principle (Baxter, 1973; Blix, 1974). This general principle has surfaced in several weapon prohibitions. First, indiscriminateness is a constitutive feature of the entire category of weapons of mass destruction (WMD). As of recently, each of these weapons – biological, chemical, and nuclear weapons – have been explicitly prohibited by a separate treaty. Second, several conventional weapons have been restricted or prohibited due to their indiscriminate effects, the treaties prohibiting anti-personnel (AP) landmines (1997) and cluster munitions (2008) being the two most recent and most prominent examples. The processes resulting in these two prohibitions function as procedural and substantial precedents for the ongoing norm-setting efforts on LAWS. In procedural terms, all three processes share their formal institutional origins in the United Nations (UN) Convention on Certain Conventional Weapons (CCW), and all were championed by NGO coalitions. In the cases of AP landmines and cluster munitions, the CCW’s failure to reach an agreement provoked eventually successful processes conducted by like-minded states outside the UN framework. The issue of LAWS initially gained traction within the UN framework in the Human Rights Council (HRC); it then moved to the CCW, where it has been debated since 2014, first in informal talks, and, since 2016, in a group of governmental experts (GGE), which used to spend 2 weeks’ time on the issue but has reduced the allotted time to 7 days in 2019. Yet, due to the lack of progress and the more or less open resistance to any regulation attempt by some major states, leaving the CCW is yet again being discussed. What is of more interest to us, though, is the substantial impact of previous ban campaigns on the framing of LAWS. The campaign against AP landmines succeeded in achieving the first complete ban on a conventional weapon by coining the image of AP landmines as ‘indiscriminate, delayed-action weapons that cannot distinguish between a soldier and an innocent civilian’ (Price, 1998, p. 628). Some years later, the ban on cluster bombs was grafted onto this existing stigma by drawing an analogy between landmines and unexploded submunitions killing civilians long after the end of conflicts (Petrova, 2016; Rosert, 2019). As mentioned at the outset of this article, the legal argument against LAWS is more complex and also involves issues such as accountability and precautions in attack. Nevertheless, the frame of ‘indiscriminateness’, which has worked out well twice in the past, has been salient since the earliest warnings against LAWS and remains a focal point of the ongoing pro-ban discourse, especially in communication from the international Campaign to Stop Killer Robots. Shortly after its formation in 2009, the International Committee for Robot Arms Control (ICRAC) announced in the first sentence of its foundational ‘Berlin Statement’ that such weapons systems ‘pose [pressing dangers] to peace and international security and to civilians’. When Human Rights Watch (HRW) embarked on the issue in 2011, the question most interesting to them was whether LAWS were ‘inherently indiscriminate’; when Article36 – an NGO advocating humanitarian disarmament, with civilian protection at its core – became another champion of a ban, the link between LAWS and civilian harm was further strengthened (Carpenter, 2014). The then UN Special Rapporteur on extrajudicial, summary, or arbitrary executions, Christof Heyns, emphasized the ‘specific importance’ of the ‘rules of distinction and proportionality’, and pointed out that the ability of LAWS to ‘operate according to these rules’ will likely be impeded (Heyns, 2013, pp. 12–13). Launched in fall 2012, the Campaign to Stop Killer Robots coordinated by HRW also placed special emphasis on the p