Today, armed drones or at least weapons-capable drones are part of states’ military arsenal quasi by default. From time to time, one hears complaints that the ethical debate about these technologies in relation to the German armed forces (Bundeswehr) lasted too long and was therefore typical of the overly concerned culture of discussion in Germany. The “drone debate” did indeed extend over a period of almost ten years, from around 2012 to 2021.[1] The first edition of “Ethics and Armed Forces” (2014)[2] is a revealing document from the initial period of this controversy. Even at that time it considered the issues of remote-controlled aerial vehicles and so-called “autonomous weapons systems” as a joint issue, despite politicians always maintaining that only drones – but under no circ*mstances “killer robots” – would be procured for the Bundeswehr.[3]
From a technological and military point of view, it is not really appropriate to conflate remote control – where humans retain complete control but are not exposed to combat situations – and “autonomy” – where the machine (supposedly!) makes its own decisions. Yet there are perspectives from which it is not entirely erroneous to examine the various new (“emerging”) military technologies from a joint perspective.[4] To a certain degree, drones and autonomous systems are developed for common goals: military action with the greatest possible precision. In the first years of this century, people talked about “surgical strikes”, meaning that the aim was to hit only the “legitimate” target as precisely as possible, while keeping collateral damage to a minimum.[5] Of course this was related to the nature of the conflicts at the time: These were primarily peacekeeping and stability operations, in which large military powers such as the United States and its allies took action against insurgents or terrorists (the Taliban) in much less developed theaters of conflict (Afghanistan). In the broadest sense, one can speak of “humanitarian interventions”, and in such deployment scenarios it is particularly important to win the “hearts and minds” of local people – or at least not permanently alienate them. As we all know, Western plans in Afghanistan did not work out, and there is a need to investigate how the use of certain technologies may have contributed to this failure. Russia’s attack on Ukraine in 2022 has in some ways brought interstate war back into our ethical debates, which gives us a slightly different mental starting point: We are once again talking more clearly about self-defense, and we know that any restriction of the methods and means of warfare could ultimately lead to the war being lost – with all the consequences. Western armed forces did not face such an intensity in their overseas deployments: They could pull out. Now the local population is bearing the consequences.
(De-)legitimization of force instead of (de-)legitimization of technology
The use of technologies – including military technologies – always takes place in a social context. It is hardly possible to make a fair assessment without reference to the context. Furthermore, it is rarely the technology itself that raises the biggest ethical quandaries, but rather the actions that are associated with this technology or suggested by it. It was not wrong to discuss the practice of “targeted killing” in connection with armed drones,[6] even though targeted killings can also be carried out with weapons (systems) other than drones. Reaper and Predator drones[7] (in conjunction with Hellfire missiles) are very well suited for targeted killing activities – which conversely also implies that the availability of the means makes this use more likely. If your only tool is a hammer, then every problem looks like a nail, Abraham Maslow noted decades ago.[8] For those who possess armed drones, the question of targeted killing is different than for those who don’t have access to this technology. This is not changed in any way by the quite correct restriction that drones should only be used in accordance with International Humanitarian Law – which does not rule out all targeted killings, and it certainly does not rule out the use of drones in principle. Therefore, justification for using this technology must be linked to the question of the justification of military force per se. The same applies to so-called “autonomous weapons systems”. Attempts have been made to rule out lethal autonomous weapons systems (LAWS) in principle by arguing that their use always violates human dignity. But this argument is perhaps more rhetorical than substantial. After all, the human dignity invoked here cannot be just a word. To serve the argument, it must become a concept. As such, however, the term requires a definition, and it is more than questionable whether a definition of human dignity that would allow such concrete normative conclusions can be universally agreed upon.[9] On the other hand, it is not questionable that these are a means of violence, and that the use of violent means requires a special kind of justification (unless violent action is ruled out in principle).
Defining the term “(lethal) autonomous weapons systems” itself presents us with great difficulties. When it is said that autonomous weapons systems “independently make life-or-death decisions”, this is an anthropomorphism, because our knowledge of what decision-making actually means comes mainly from our self-experience in our practice of deciding. Nevertheless, this metaphorical use of language is understandable, as we ourselves seem to delegate our decision-making capability to the machine when using such a system. This is not always unreasonable, as we know that we cannot keep up with the highly developed machine either in terms of sensor(y) systems – i.e. recognizing relevant situational factors – or in terms of computing power, for example when assessing the consequences of actions. We therefore take ourselves out of the “decision loop” as it were, and hope for the technology to make the “better” and – very important in military conflicts – faster decision. But the fact that we try to withdraw (and remain responsible for the withdrawal) does not yet turn the technical instrument into a moral actor. “Autonomous weapons systems” are at best still controlled by what is referred to as “weak artificial intelligence”. If they were conscious and therefore had a moral awareness, they would be a “strong artificial intelligence”, which has not yet been – and may never be – invented.
Autonomous weapons systems can be deployed in “swarms”, which also makes them a means of mass destruction. Yet the finely tuned technology with its excellent sensor systems and computing capacity is supposed to prove its worth in particular where finely focused, small-scale effects are required, i.e. where precision matters. Autonomous weapons systems in some ways represent an advancement in remote-controlled systems (like drones), and thus are designed for targeted actions – such as targeted killings or “signature strikes”, where the target remains anonymous but is identified by patterns of behavior. The devices or systems (of linked devices, for example) are becoming ever smaller as technology progresses. Currently they have shrunk to the size of the “Black Hornet”[10] used by the German armed forces, i.e. approx. 12 to 16cm in length. But further miniaturization is surely feasible. Combined with effective facial recognition, a completely new type of targeted killer could emerge; a kind of “autonomous weapons insect”. The idea of such devices becoming available to private individuals and on a wide market is quite frightening.
Nevertheless, we can certainly think of cases in which we consider the use of such instruments to be morally acceptable, for example in the fight against terrorists or crimes of aggression. But what normative set of rules should apply here and contain their use? Based on the human rights protection regime, the prevention of immediate danger is certainly permissible. But what if the attacker or attackers are only in the process of preparation, and the danger is not immediate in the strict sense? In this case, human rights law does not permit a lethal attack on a person. International humanitarian law, on the other hand, does permit the killing of enemies (combatants) even without an immediate threat. But a certain amount of violent action on both sides is required in order to be able to speak of an armed conflict, whereupon this legal regime applies. Michael Walzer is a social philosopher who has written one of the most influential books on the ethics of war. In the wake of the terror attacks at the beginning of this millennium, the problem became quite clear to him: Against certain terrorist opponents, a state is rather limited in its permissible defenses if it is strictly bound by human rights norms, with the result that it may not be able to mount an effective defense. Yet in order to qualify a conflict as war, in which combatants can then legitimately use military force against enemy combatants (as Walzer sees it[11]), the fighting would have to reach a threshold of violence that terrorism does not reach, for example. For this reason, he argues that in addition to a jus ad bellum [right to war], states also need a jus ad vim [right to limited force] so that they do not have to remain inactive in violent situations below the threshold of war.[12]
The construct of jus ad vim
Michael Walzer himself preferred not to elaborate such a jus ad vim, but the suggestion he put forward was taken up by others. In 2013, Megan Braun and Daniel R. Brunstetter published two essays discussing the use of armed drones and the question of jus ad vim.[13] Following criticism of the model,[14] Brunstetter expanded his ideas and published them as a book a few years ago.[15] The book title Just and Unjust Uses of Limited Force. A Moral Argument with Contemporary Illustrations is very self-assured, as it alludes to Michael Walzer’s classic, whose subtitle is A Moral Argument with Historical Illustrations. The apparent implication is that Walzer worked with examples from history, and so jus ad bellum is largely a matter of history. Brunstetter, on the other hand, has developed and elaborated a theory of the present using contemporary examples. – The Russian war of aggression against Ukraine has shown that there can still be much in the present that we may have thought was consigned to history. It is bold of Brunstetter to borrow from Walzer in this way, but it does not lessen the value of reflection on violence below the threshold of war. Especially as Brunstetter uses Walzer’s method and seeks to identify the moral concepts that permeate actual political action from the way in which this action is talked about.
This debate about a jus ad vim has not yet reached Germany, which could be largely due to people’s reluctance to discuss the Bundeswehr as an offensive actor. Even in the debate about acquiring weapons-capable drones, the notion that they would protect soldiers was invoked time and again. Now, protection arises primarily from the drones’ reconnaissance capabilities. If you arm them, then you also want to hit enemies. But this is largely being downplayed. We also hear little in public about the activities of the German Special Forces Command (KSK) – activities that can certainly be in the realm of force short of war. The Bundeswehr website merely states:
“The capabilities of the Special Forces Command expand the options for action available to the Federal Republic of Germany across the full spectrum of the armed forces’ tasks. Special forces are particularly suited to operations that are carried out in a controlled manner and can be limited at any time.”[16]
“Controlled and limitable” are also the key terms for force which – so it is claimed – is to be regulated by a jus ad vim. As his initial example, Brunstetter cites limited air strikes in the Syrian civil war that were carried out in retaliation for the use of chemical weapons in 2017 and 2018.[17] This was not a full participation in the war, nor was it aimed at bringing about “regime change” in Syria. It was about sending a signal that the use of banned nuclear, biological or chemical (NBC) weapons will not be tolerated. Recently, there have been reports of military strikes against the Houthi rebels, who are a party to the civil war in Yemen. In conjunction with the multinational “Operation Prosperity Guardian”, from December 2023, similar to the EU “Aspides” mission, strikes were carried out against positions held by this group – not with the primary aim of intervening in the Yemeni civil war, but to prevent further attacks on civilian shipping in the Red Sea and Gulf of Aden. Remote-controlled drones are used in such attacks – and perhaps soon “autonomous” drones too.[18] It can be assumed that there was no jus ad bellum for a full-scale war in Syria, nor is there currently one in the Red Sea. But does this mean that no military intervention is possible at all?
Special forces are just one case of the use of “limited force”, which Brunstetter believes should be regulated by a model of jus ad vim. Along with drone strikes, he also includes no-fly zones. This is worth mentioning in our context because the question of suitable technology arises particularly in the case of control regimes such as no-fly zones. Since a fast combination of reconnaissance and strike capability is required, this would certainly be one possible area of application for (AI-assisted) autonomous weapons systems. The ethical question of whether autonomous weapons systems may be used in this scenario cannot be answered on the basis of the technology itself, but with reference to the context of use. What type of use is envisaged, and with what goals? No-fly zones can serve to protect civilians. The technology would then simply help to ensure that this goal can be achieved, for example by immediately detecting and “neutralizing” any ascending flying object (manned or unmanned) of the aggressor’s.
The idea is to potentially apply jus ad vim in the intermediate area between human-rights-compliant law enforcement, as is possible and required in functioning constitutional states (law enforcement paradigm), and the realm of full-scale war, where human rights play a subordinate role and International Humanitarian Law provides the relevant legal guidelines – including the permission to attack and strike enemy combatants.[19] In contrast to the so-called “revisionist theory of just war”, which attempts to extend our basic concepts of human rights into the field of war, jus ad vim comes more from the opposite side. Brunstetter explicitly states that he wants to defend “the recalibration of just war principles, combined with new principles that are specific to contexts of limited force”.[20] Accordingly, jus ad vim is not about new or different principles – except for one – other than those already known from the “just war theory”: “just cause, legitimate authority, right intention, necessity or last resort, proportionality and reasonable hope of success”, as Jai Galliott puts it in the introduction to the first comprehensive anthology on jus ad vim.[21] The phrasing here is not innocuous, indeed it already sets the tone: In contrast to Thomas Aquinas (Summa Theologiae II-II, question 40), just cause is mentioned here before authority, the last resort criterion is also understood as necessity, and instead of “reasonable prospect of success” it says “hope of success”, which somewhat softens this criterion by subjectivizing it. All these subtleties, which on closer inspection are more than just subtleties, could be discussed at length. Here we can only briefly touch on the problems in the relationship between just cause and legitimate authority. The traditional approach in considering the conditions under which war can be lawful (“utrum bellum sit licitum”, Thomas Aquinas; not whether it can be “just”) understands authority to be that of the (Christian?) sovereign. In modern times this has been transferred to the state. The other examples given by Brunstetter and others for jus ad vim also envisage the actors to be states. The question of whether as such they are legitimized to use violence at all takes a back seat to the question of the reason for the violence. But this is by no means self-evident. One could certainly argue that other political groups may also have a right to violence. For example, separatists who perceive the state as foreign rule and want a separation from the old state and a new state of their own. Particularly in this respect, so-called “revisionist” approaches in the current debate on “just war” – such as that of Cécile Fabre[22] – are proving to be much more efficient. This is because “revisionism” begins very fundamentally with the original authority of the individual, and every authority of groups or states has its normative source in each of the individuals.[23]
A right of the strong?
On the other hand, this revisionist approach is certainly problematic with regard to limiting violence, as most people would probably prefer to see the use of emerging military technologies concentrated only in the hands of states. Although many states give cause for concern through their possession of drones and increasingly AI-controlled weapons systems, non-state groups are likely to be even more worrying as owners and users of remote-controlled, autonomous or semi-autonomous technologies – and they (these groups) are more difficult to accommodate within the framework of international law. But here jus ad vim also offers few starting points for containment. It seems to be very much based on the de facto balance of power during the noughties of this century: At that time, the use of drones was mainly a matter for states, and the number of states initially seemed to be limited to those whose fundamental authority was not in question – such as the United States of America or the United Kingdom. This led to the not entirely unjustified suspicion that jus ad vim was primarily a right of the strong. There is a Bavarian saying, “Wer ko, der ko” [Those who are able to do it have a right to do it] – this may often seem to be true in fact, but can serve as a normative model only with cynicism. Brunstetter and other proponents of this approach are aware of the objection and try to forestall it by adding on a criterion of “probability of escalation”[24] – as was briefly alluded to above. The violence justified by the jus ad vim must always be limited in such a way that it does not escalate into a “full” war: “Vim acts can be justified only if there is a reasonably low risk that they will result in the escalation of violence to war.”[25] Elsewhere, this “principle” is supplemented by the demand for a “predisposition toward maximal restraint”.[26]
But how can this principle limit anything at all?[27] The term “escalation” remains undefined. One gets the impression that escalation would occur primarily if a war between two states were to result from attacks by a state party on a non-state party to the conflict. That would be a useful definition. It is a similar situation with “maximal restraint”: Maximal restraint would be the renunciation of violence, however this is not what is being demanded. So how do we measure whether the restraint is at a maximum? Apparently, it’s about a set of relations with regard to a fixed goal: We want to defeat A, but with as little violence as possible. That sounds more moral than it is, because the goal of victory is maintained. It could also be said the other way round: We will use whatever violence is necessary to win. Therefore, jus ad vim does not present us with a deontological standard that could also mean that we have to sacrifice ourselves (or at least accept disadvantage) in order to uphold the standard. Instead, it instrumentalizes moral norms toward goals of actions that are not considered to require any further justification.
Persistent conflict, fragile peace
Of course, it must be admitted that the concept of victory is rightly problematized in this context. Situations in which jus ad vim is brought into consideration are obviously not about using force to make an enemy surrender. The aim is merely to achieve a “truncated victory”. The opponent who is fought with vis can survive (politically), but for example must not continue certain practices (such as the use of poison gas). It is a benefit that the debate on jus ad vim openly addresses the difficulty of finding an appropriate concept of victory today – even in view of existing military technologies.[28] On the other hand, being satisfied with a truncated victory is not praiseworthy for its modesty in every respect. These “incomplete” victories leave a latent conflict situation that can degenerate into violence. Again, the situation can be illustrated by drones and autonomous weapons systems: Truncated victory also means that the drone bases remain in place and the surveillance flights continue. The next non-compliant incident could result in another military strike. Those in possession of the required technical equipment will find this situation easier to come to terms with than those who have become the “object” of this technology, as it were. Peace in this case can no longer be understood as anything other than fragile. The fact that Brunstetter wants to start with jus post vim [justice after limited force] does little to change this, as jus post vim dispenses with “classic” criteria of jus post bellum such as “rehabilitation, regime change, war crimes trials, and so on”.[29] They are not feasible in such a context, so he says[30] These dropped criteria obviously also include reconciliation. Yet without reconciliation, there is no positive peace.[31] Violence is perpetuated at least as latent violence.
In the meantime, reality seems to have somewhat escaped this theoretical construct – if it can be described as such at all. Israel’s ongoing (as of May 2024) military deployment in the Gaza Strip following the devastating Hamas attack on October 7, 2023 is in some ways not a full-scale war like the one in Ukraine. In Ukraine, two state armies and state parties to the conflict are fighting each other; in Gaza, a state army is pursuing a political group and its armed and extremely violent members. The “war” in the Gaza Strip is an asymmetrical conflict, while the one in Ukraine is at least formally symmetrical. Nevertheless, the Israeli military operation goes far beyond drone strikes or other isolated acts of violence, and the declared goal is the destruction of the enemy.[32] This hardly fits with the concept of jus ad vim. So is there a need for a new category of a jus ad magnam vim [right to great force] for such acts of extreme violence? Should this also be an independent theoretical construct that is separate from jus ad bellum and jus ad vim, as it is from law enforcement? We thus arrive back at the initial situation that motivated David Rodin’s[33] and Jeff McMahan’s[34] criticism of Michael Walzer’s theoretical separation of jus ad bellum and jus in bello: It is not plausible to assume special ethics such as an ethics of war for more and more social fields. Even the norms of violence and conflict must be compatible with our everyday moral assumptions. Thus the approach based on the violation of the individual’s rights presents a much more convincing alternative for the normativity of violent action than the distinction of special social fields. As a deontological ethics, it is focused above all on limiting violence, even if compliance with the norm is associated with costs. To feel bound by norms in the sense of “just war” also means having to accept a possible disadvantage, since the norms for violence do not apply with reservations.
What does this mean for the new military technologies mentioned at the outset? A “maximum restraint” that would go beyond Brunstetter’s approach and comprise only defense against direct aggression (while largely refraining from counter-strikes) seems more necessary than ever. Admittedly, this runs counter to the temptation to exploit a military technology advantage to weaken the enemy or for retaliation. Nevertheless, technologically superior states should take such asymmetrical constellations as a particular opportunity to bring to their mind the contribution that increasingly “dehumanized” technologies make to a latent permanent conflict. It is a different matter to use these technologies to help reduce one’s own vulnerability than to rely on their supposed suitability for a precise, limited use of force within the framework of jus ad vim. In the latter case, the means now available – such as autonomous weapons systems – could make us all losers.
[1] A good overview is provided in this blog post by Schörnig, Niklas (2022): Die Heron wird bewaffnet. PRIFblog, April 6. https://blog.prif.org/2022/04/06/die-heron-wird-bewaffnet-mehr-als-10-jahre-drohnendebatte-finden-ihr-ende/. (All internet references accessed May 8, 2024.)
[2]https://www.ethikundmilitaer.de/ausgabe/2014-01/overview.
[3] On the ethical aspects, cf. Koch, Bernhard (2019): Die ethische Debatte um den Einsatz von ferngesteuerten und autonomen Waffensystemen. In: Werkner, Ines-Jacqueline and Hofheinz, Marco (eds.): Unbemannte Waffen und ihre ethische Legitimierung. Wiesbaden, pp. 13-40.
[4] Cf. Koch, Bernhard (2019), see footnote 3.
[5]https://en.wikipedia.org/wiki/Surgical_strike.
[6] Cf. Koch, Bernhard (2011): Ein Beigeschmack von Selbstjustiz. Lässt sich das gezielte Töten von Terroristen rechtfertigen? In: Herder Korrespondenz. Monatshefte für Gesellschaft und Religion 65 (7), pp. 352-356.
[7] In other words, so-called MALE drones (https://de.wikipedia.org/wiki/Medium_Altitude_Long_Endurance), which should not be confused with the much smaller single-use drones that Iran apparently supplies to Russia, which in turn uses them to attack Ukraine (https://de.wikipedia.org/wiki/HESA_Shahed_136).
[8] Cf. Maslow, Abraham (1966): The Psychology of Science. A Reconnaissance. New York/London, pp. 15 f. Cf.Koch, Bernhard (2022): Technikethik. In: Merkl, Alexander and Schlögl-Flierl, Kerstin: Moraltheologie kompakt. Grundlagen und aktuelle Herausforderungen. Regensburg, pp. 340-350.
[9] Cf. Koch, Bernhard (2024): Menschenwürde und “autonome” Robotik: Worin besteht das Problem? The article is part of this issue.
[11] So-called “revisionists” (see endnote 22 below) see at most one of the warring parties (namely the one with the just cause) as being morally justified to wage war.
[12] Walzer, Michael (2006): Just and Unjust Wars. 4th ed. New York., pp. xiv f.
[13] Braun, Megan and Brunstetter, Daniel R. (2013): State of the Union: A Decade of Armed Drones. In: Brown Journal of World Affairs 19/2, pp. 81-95; by the same authors (2013): From Jus ad Bellum to Jus ad Vim. Recalibrating our Understanding of the Moral Use of Force. In: Ethics & International Affairs 27/1, pp. 87-106.
[14] Frowe, Helen (2016): On the Redundancy of Jus ad Vim. A Response to Daniel R. Brunstetter and Megan Braun. In: Ethics & International Affairs 30/1, pp. 117-129.
[15] Brunstetter, Daniel R. (2021): Just and Unjust Uses of Limited Force. A Moral Argument with Contemporary Illustrations. Oxford.
[16] (Translated from German.) https://www.bundeswehr.de/de/organisation/heer/organisation/division-schnelle-kraefte/kommando-spezialkraefte.
[17] Cf. Brunstetter, Daniel R. (2021), see footnote 15, pp. 1f.
[19] Jai Galliott thus brings into play not only the fact of military technological development and its significance for the shaping of jus ad vim, but also the normative call for a “responsibility to protect”. However, his account too gives more weight to the technological aspect. Gilliott, Jai (2019): An Introduction to Force Short of War. In: by the same author (ed.): Force Short of War in Modern Conflict. Jus ad Vim. Edinburgh, pp. 1-12, p. 8.
[20] Brunstetter, Daniel R. (2021), see footnote 15, p. 13.
[21] Galliott, Jai (2019), see footnote 19, p. 2.
[22] Fabre, Cécile (2012): Cosmopolitan War. Oxford.
[23] Jai Galliott criticizes revisionism because making the distinction between “objectively just and unjust causes [is] nigh impossible”, which the revisionists do not adequately appreciate (Galliott, Jai (2019), see footnote 19, p.5). But if this is the case, then jus ad vim is all the more open to criticism, because it depends on a clear distinction being made in this respect. Otherwise, both sides quickly have a jus ad vim.
[24] “This criterion [the probability of escalation] blends elements of the jus ad bellum proportionality and probability of success criteria to conceive the risks of using limited force.” (Brunstetter, Daniel R. (2021), see footnote 15, p. 24).
[25] Brunstetter, Daniel R. (2021), see footnote 15, p. 24. – One of the curious things about Brunstetter’s book is that he always uses the word “vim” (in the Latin accusative) when it should really be “vis” (in the nominative).
[26] Brunstetter, Daniel R. (2021), see footnote 15, p. 17.
[27] On the criticism of this additional principle, cf. also: Lango, John (2019): Just War Theory, Armed Force Short of War and Escalation to War. In: Galliott, Jai (ed.): see footnote 19, pp. 148-169.
[28] Just as in “jus in vi” the difficulty of making a tenable distinction between enemy combatants and civilians is also addressed; in addition, it is also made explicit that “jus in vi” is not only about protecting civilians, but also about minimizing one’s own victims on the side of the state party to the conflict (cf. Gailliott, Jai (2019), see footnote 19, p. 6).
[29] Brunstetter, Daniel R. (2021), see footnote 15, p. 23.
[30] Ibid.
[31] Cf. on reconciliation Koch, Bernhard (2023): Anger and Reconciliation. In: Conatus - Journal of Philosophy8(2), pp. 279-298. doi.org/10.12681/cjp.35255.
[32] Cf. for example https://www.zeit.de/politik/ausland/2023-10/israel-ueberblick-gazastreifen-flucht-netanjahu-hamas-vernichtung. I am indebted to Rüdiger Frank for this and other references.
[33] Cf. Rodin, David (2002): War and Self-Defense. Oxford.
[34] Cf. McMahan, Jeff (2009): Killing in War. Oxford.
