Once Again: Drone Killings and International Law

[Kai Ambos is Professor of Criminal and International Law at the University of Göttingen and a judge at the Kosovo Specialist Chambers, The Hague. He writes here in his academic capacity. The author would like to thank Dr Lippold, Göttingen, for his important suggestions, and Dr Margaret Hiley for the preparation of the English version.]

Reports on the U.S. drone killings in connection with Afghanistan reveal an astonishing lack of concern regarding the legality of these operations. International law seems to play no role whatsoever in their assessment. Its marginalization, already lamented previously in connection with the Soleimani killing (see here in German and Spanish, with further references), continues to progress, despite the fact that Germany and other Western nations see themselves protagonists of international law. Indeed, the permissibility of such killings under international law is anything but clear.

In the view of the U.S., these operations are covered by the so-called war on terror against the IS and related groups. However, it is questionable whether this actually constitutes a “war” in the first place, in the legal sense of the term, namely an “armed conflict” pursuant to international humanitarian law (IHL) or law of armed conflict. This is because the “war on terror” involves military action not against another State, but against a non-State actor that moreover acts transnationally and as a loose network. In principle, non-State actors acquire (partial) subjectivity under international law only when they exercise a certain territorial rule and thus become a de facto regime. Traditionally, this was assumed, for example, in the case of colonial liberation movements that had gained control of a given territory; it also applies to local insurgency movements that, like the Taliban, pursue territorially limited interests and exercise limited territorial rule before actually seizing power.

A non-international conflict may also arise between such movements and the government concerned if – within the meaning of Common Article 3 of the Geneva Conventions and Article 1(1) Additional Protocol (AP) II – the conflict reaches a certain intensity and the non-State actor is sufficiently organised to be considered a party to the conflict. (Cf. ICRC Commentary, para. 456 ff.) In the case of terrorist attacks, it is debatable whether the required intensity threshold is met, and in the case of a transnational terrorist network, doubts may arise as to whether it possesses a sufficient degree of organization, precisely because of its lack of a territorial nexus, let alone a territorial consolidation. Even though territorial control on the part of the non-State actor is not required in Common Article 3 (Article 1(1) AP II thus sets higher requirements in this regard), according to both provisions a non-international conflict must take place on the territory of a High Contracting Party; the borderless expansion of such a conflict – in the sense of a worldwide war on terror that completely abandons the territorial nexus – is thus not covered by IHL (see here and ICRC Commentary, para. 516). Whether a terrorist network – detached from a specific territory – can possess general subjectivity under international law, which would include the possibility of attacking it, is a difficult question. In any case, States have an inherent right to self-defence, which can also be directed against non-State actors (BVerfG, 2 BvE 2/16, paras. 50–51).

At any rate, such extraterritorial operations may constitute a violation of the sovereignty of the State of sojourn. If this State – for example, Afghanistan (now represented by the Taliban) – did not consent to such an attack, the strike would constitute a violation of the principle of non-intervention derived from the principle of the sovereign equality of States (Article 2(1) UN Charter) (cf. ICJ, Nicaragua Judg. 1986, para. 202); it could also amount to a violation of the prohibition of the use of force (Article 2(4) UN Charter). It is noteworthy in this context that Afghan sovereignty is explicitly recognised in both the NATO Statute of Forces Agreement (Art. 4(2)) as well as in the bilateral agreements between the USA and Afghanistan (Article 3(2)).

Note, however, that the right to self-defence is affirmed in the same sentence, and could indeed be invoked by the attacking State (the U.S.) if the conduct of the non-State actor can be attributed to the State of sojourn (Afghanistan), either pursuant to the law on state responsibility because of effective control, or if the State of sojourn offers the terrorist actor protection and is unwilling or unable to take effective action against it (cf. once again BVerfG, 2 BvE 2/16, paras. 50–51). In this case, an international armed conflict may even arise between the attacking State and the State of sojourn (ICRC Commentary, para. 511). A prerequisite for invoking self-defence, however, is that the use of force against which the attacking State is defending itself exceeds the threshold of an “armed attack” within the meaning of Article 51 of the UN Charter. Precisely this was the case with the first Taliban government and the attacks of 11 September 2001, which were planned in and directed from Afghanistan (under Taliban shelter). The UN Security Council took this into account in several resolutions, particularly in Resolution 1373/2001, which in para. 2(b) authorises “necessary steps to prevent the commission of terrorist acts” (in greater detail here). Accordingly, at that time, the U.S. government had good reason to invoke its right of self-defence. Yet, today’s situation is not comparable because, as far as is known, the Taliban do not support or give shelter to “ISIS-K”.

If one nevertheless assumes the possibility of a “war” against ISIS in the legal sense, this does not answer the question of exactly which persons may be attacked. Each member of the network? Only certain leaders and the actual fighters? The first-mentioned broad view is espoused by the so-called membership approach, which, however, needs to develop convincing criteria of “membership” and prove that they have been met in individual cases. The fact that IHL only deprives civilians or (formal) non-combatants of immunity from attack if they “take a direct part in hostilities” (Article 51(3) AP I and Article 13(3) AP II) speaks in favour of limiting attacks to specific individuals. In this case, one would need to clarify what is understood by such participation, and whether the potential target of an attack has indeed participated in this way. A distinction would have to be made between leaders, direct combatants (e.g. suicide bombers), and mere supporters. If, in addition, there are civilian casualties (“collateral damage”), the difficult question arises as to when this “collateral damage” is disproportionate (cf. ICRC Commentary,paras. 514–515).

If, by contrast, the existence of an armed conflict is denied, then the permissibility of the killing is ruled by peacetime international law. In line with domestic regulations, this law permits killings only in extreme exceptional cases, such as in the case of self-defence. Otherwise, alleged terrorists, like all criminals, must be prosecuted and judged according to the rule of law, which in particular means granting them a fair trial before an independent and impartial tribunal. While the death penalty is possible beyond the scope of the European Convention on Human Rights, its application still requires a fair adjudication and cannot be based on the mere suspicion of an intelligence agency. Indeed, the already mentioned antiterrorism Resolution 1373/2001, drafted with heavy involvement of the United States, unambiguously formulates in para. 2(e) that participants in acts of terrorism must be brought to justice according to the rule of law, not liquidated on suspicion.

In a nutshell, the decision to carry out such killing operations is based on opaque intelligence information that can be trusted – or not. Such pre-emptive strikes are always only verifiable ex post facto; if the suspicion proves unfounded, this comes too late for the victim of the attack. As a rule, no independent official review is carried out following the attack; there may be journalistic or other private investigations, and, in exceptional cases, perhaps also an investigation under (international) criminal law. Ultimately, these attacks are executive military liquidation decisions based on political ascriptions and personalisations (“terrorist”), as already criticised here, beyond any democratic or judicial control.