The international law framework regulating the use of armed drones – Heyns, Akande, Hill-Cawthorne and Chengeta

In our first review, we have looked at  ‘The international law framework regulating the use of armed drones’ by Christof Heyns, Dapo Akande, Lawrence Hill-Cawthorne and Thompson Chengeta in International and Comparative Law Quarterly vol 65, October 2016 pp 791–827.

It is currently available for free from CUP. See here.

The article was recently presented at the International and Comparative Law Quarterly annual lecture at BIICL by Dapo Akande and Lawrence Hil-Cawthorne. It was a great event and generated some fascinating questions and discussion.

The article is a greatly expanded version of a report on presented to the UN in his former role as Special Rapporteur on extrajudicial, summary or arbitrary executions. His 2013 report is available here.

It also builds upon the BIICL event on drones last year. The event summary for which is here.

This post is the first for our research network. We are hoping to generate ideas, comments, and hopefully collaboration between group members and beyond. Please feel free to use the comment box below, or if you would like to get involved in the network, send an email to

The article is rather long, and network member Maziar Homayounnejad (KCL) has written a summary of the article. Following the summary will be a few thoughts on the article from network members.


In their article on the regulation of the use of armed drones, Heyns et al. put forward a holistic framework encompassing three distinct branches of international law (IL).

  • Jus ad bellum, which regulates the resort to military force, only allowing this where there is consent from the territorial State; in lawful self-defence against an actual or imminent armed attack; or pursuant to Security Council enforcement action.
  • Jus in bello, also known as international humanitarian law (IHL) and the law of armed conflict (LoAC). This regulates the actual conduct of hostilities and offers expanded (belligerent) rights to kill without warning, based on enemy status and civilian conduct. It also permits a ‘non-excessive’ number of civilian deaths, so long as these are ‘collateral’ and not directly targeted.
  • International human rights law (IHRL), which applies during peacetime to regulate police and law enforcement IHRL restricts the right to use lethal force to where it is absolutely necessary to protect against an imminent threat to life; and proportionate in that it is the bare minimum force required to achieve its legitimate aim and strictly tailored to circumstances.

The authors argue that armed drones are different to traditional weapons of war in that they are used to inflict low-level violence, drawn out over a long time period and often spanning several countries (including those which are far away from any contact zone). Accordingly, legal analysis of drone strikes cannot be limited to any one branch of IL, but must be shown to satisfy all three; the requirement of holistic application of IL is a constant theme to which the authors keep returning. Not least because States should be prevented from cherry-picking the framework in which they can demonstrate legal compliance, while ignoring the ones they may be violating. Preventing such tendencies has an important role in preserving the exceptional nature of war and, consequently, the right to life (RtL), which is more effectively protected under IHRL.

Equally important is to avoid confusing and conflating common terms, which may appear under two or more legal frameworks but have vastly different meanings under each one – ‘imminence’ and ‘proportionality’ being two such examples. Again, this prevents both cherry-picking and inappropriate increases in the incidence of war and loss of life.

After introducing the applicable legal frameworks and the relationships between them, the authors analyse in some detail jus ad bellum, IHL, and IHRL. Three particularly interesting points stand out from this. The first relates to the Tadic criteria for political violence qualifying as a non-international armed conflict (NIAC); namely, a) protracted armed violence and b) organisation of Parties. This has come under strain as far as individual drone strikes are concerned, in that it is doubtful that such isolated acts (if not in the context of a pre-existing armed conflict) will meet the required threshold of violence to invoke belligerent rights under IHL. If so, then the more restrictive IHRL regime will apply exclusively because an individual drone strike is unlikely to constitute protracted armed violence (unless a number of strikes can be aggregated as a ‘workaround’ to meet this threshold). By contrast, it is worth noting that autonomous weapon systems (AWS) are expected to pose challenges in relation to the second criterion (organisation of parties). This is because autonomous capabilities are more likely to enable protracted armed violence to occur with a relatively small number of persons and a looser organisational structure[1]. Of course, AWS used for targeted killings may also raise challenges in relation to the required threshold of violence, just as remotely-piloted drones do; this is worth bearing in mind as drones become progressively more autonomous over time.

A second and complementary point that the authors make relates to the expansive interpretation of the applicability of IHRL. The traditional approach to human rights required a clear geographical nexus, hence the mantra ‘on your territory and subject to your jurisdiction’. Now (and certainly over the past 20 years or so) IHRL has evolved and extended into areas of armed conflict, where courts have reconsidered the proper scope of governmental authority, and have deemed certain aspects of a hostile environment (e.g. occupation, detention, etc.) as falling within that scope. In addition, there is now a ‘divide and tailor’ principle, whereby the State’s positive obligation to protect and enforce human rights is separated from its negative obligation to refrain from arbitrarily depriving someone of their rights. The authors argue that the former requires territorial control and ‘jurisdiction’, but the latter merely requires ‘authority or control’ over the specific individual concerned. By definition, this invokes IHRL whenever a terrorist suspect is unilaterally pursued and lethally targeted abroad. Hence, the authors describe IHRL as the ‘default’ regime in such circumstances, whether or not an armed conflict exists; whether or not the NIAC threshold is satisfied. Of course, where there is no NIAC or other armed conflict, IHRL applies exclusively to impose both positive and negative HR obligations, both in relation to those targeted and those who may be caught in the crossfire. Where a NIAC does exist, IHRL applies in limited form to impose only the negative obligations on State action.

Thirdly, and again complementary to the above, the authors warn against any liberal aggregation of violence, such that we easily and frequently arrive at the conclusion that a NIAC exists. This is because when IHL is invoked, not only do less protective rules apply to those targeted, but it also means unsuspecting communities far away from any contact zone may lawfully be ‘significantly’ affected; indeed, so long as collateral deaths are not ‘excessive’, the IHL principle of proportionality allows them. This concern is echoed by other authors (such as McDonald, 2015), who point out that drone strikes in areas that are putatively in peacetime effectively deprive unsuspecting civilians of the option of seeking refuge abroad, in order to separate themselves from violence that would otherwise kill them. Not to say that being a refugee is desirable, but if the choice is between a) refugee status and b) the risk of life-changing injury or death, most civilians would likely opt for the former. Yet, drone strikes far away from an established contact zone effectively deprive them of that choice. Thus, in ambiguous circumstances we should not be quick to ‘aggregate violence’ in order to allow State action to benefit from IHL belligerent rights.

The authors sum up by reminding us of three things. Firstly, the long-standing legal framework for maintaining international peace and the RtL is both coherent and well-established, and has stood the test of time. Yet, States are using drones to exercise a police / law enforcement function to counter potential threats from allegedly new forms of conflict, which originate abroad; this poses a danger to the protection of life because the usual tools of policing (e.g. capture) are not available and the more permissive IHL framework is often relied upon instead. However, the long-standing framework mentioned above need not and should not be abandoned on account of terrorism and ‘new’ forms of conflict. On the contrary, the framework and its standards should be more diligently applied, especially in light of the increasing number of States accessing this technology and putting further pressure on the RtL.

Secondly, to reiterate, the legal paradigm applied to drone strikes makes an important difference to its permissiveness, hence an assumption of global NIAC makes it easier to justify lethal force than if IHRL becomes the dominant legal framework. Yet, the global NIAC concept is controversial: even if accepted in principle, it must be assessed against objective legal criteria, not subjective self-interest; even if it applies in a given situation, it should not be assumed to modify IHRL, even less so to displace it. As an afterthought, the authors remind us that IHL permissibility does not automatically mean desirability or wisdom: to protect the RtL, it is important to maintain the idea that force remains the exception. This is especially true for unsuspecting civilians located far away from any contact zone, who may find themselves incidentally injured without prior warning or the effective choice to seek refuge.

Finally, considering the controversy and debate surrounding the above, the authors conclude there is “an urgent need for the international community to gain greater consensus on the interpretation of the constraints that IL in all its manifestations places on the use of drones.” This is important not just for those being targeted, but also for the unsuspecting civilians mentioned above; and it is, more broadly, in the interests of keeping intact a viable and strong system of international security where the rule of law is a central feature. Drones should follow the law, not vice versa.

To begin the search for greater consensus on the regulation of armed drones, the authors propose seven fundamental principles:

  1. The current IL framework is adequate to govern drone strikes;
  2. The RtL can only be adequately protected if all IL constraints on use of lethal force (UoLF) abroad are complied with holistically; that is, individual and cumulative compliance with jus ad bellum, IHL (where applicable) and IHRL;
  3. IL norms on UoLF must not be abandoned or modified to suit the current use of drones;
  4. Any extra-territorial drone strike must comply with jus ad bellum;
  5. Where a drone strike takes place within the narrow confines of international armed conflict or NIAC (a factual question governed by objective legal criteria), it is governed by both IHL and IHRL;
  6. Outside those narrow confines, any killing is governed solely by IHRL and must meet the relatively stricter necessity and proportionality requirements developed in that jurisprudence; and
  7. There should be transparency surrounding all drone operations to enhance accountability.

[1] Stuart Russell, a leading computer scientist, illustrates the point with a (perhaps dramatised) example:

“A million Kalashnikovs can kill an awful lot of people, but only if carried by a million soldiers, who require a huge military-industrial complex to support them…

A million autonomous weapons, on the other hand, need just a few people to acquire and program them – no human pilots, no support personnel, no medical corps. Such devices will form a new, scalable class of weapons…they tip the balance of power away from legitimate states and towards terrorists, criminal organizations, and other non-State actors.”

While Russell’s comment may sound like a distant and dystopian future, modern developments in swarm technologies are certainly moving in that direction, with the US Navy’s LOCUST programme running successful trials of eight-drone swarms controlled by a single person penetrating sophisticated ship defences. The aim is to increase this to 50-drone swarms controlled by a single operator, all at a lower cost than a single Harpoon anti-ship missile. Should these Navy plans prove successful, they will enable significant violence to be inflicted with progressively less need for command and control of the battlespace. Furthermore, the miniaturisation involved will increase the likelihood of capture, reverse-engineering and proliferation to non-State actors. Accordingly, it may become more difficult as a matter of law for non-State actors acquiring these technologies to qualify as an ‘organised’ armed group.


Now, we move onto comments on the pieces from network members.

Mike Ryder, Lancaster University

Very interesting for me, as a non-Law student, is the question here of the obligation and responsibilities of the state in which ‘terrorists’ (be they criminals/enemies) are taking refuge.According to the authors, states are bound to their own human rights obligations preventing violations to the right to life of those within its borders (822), but there surely then remains the question of the state’s role in also policing itself internally and protecting the rights of others both internally and among the wider global population. Should the USA (for example) expect such states to assist in the tracking and detainment of suspected terrorists, or is the use of drones a signal that the normal situation of international law has broken down; that the state in question is incapable of policing its population effectively, and as such it is for others to take the law into their own hands?Does this then mean that the states in question, housing these ‘terrorists’ are not really true states at all? (NB: I used inverted commas for their guilt is not clear.) 

The other question here that arises for me is the question of the enemy vs the criminal.According to this paper, terrorists are ‘non international’ actors, so the conflict with them is a ‘non-international armed conflict’ (805). However for this to be the case, the enemy group must be organised: it requires a headquarters, strategy, and single voice (808). How then can any conflict with terrorists be ‘non-international armed conflict’ if often the success of terrorist organisations based on their diffuse, non-crystallisable nature? For me, this position seems to fail to understand the nature of the modern terrorist.


Joshua Hughes, Lancaster University

This article by Heyns et al is a great overview of the current legal framework governing drones, and notes a large number of issues that are being dealt with, and need to be by the international community. However, as Sir Michael Wood questioned at the ICLQ lecture, why just focus on drones? Most of this assessment could be applied to special forces counterterrorism raids.

As the authors note in the article (and restated by Akande at the ICLQ lecture about the same article), the biggest issue in relation to use of armed drones in targeted killing is the blurring of the line between war and peace, potentially resulting in a perpetual war (p.792). Which of course is a result of the current political situation where NATO states do not wish to openly risk the lives of their soldiers in potentially politically damaging operations in the middle-east. However, questions about the potential use by law enforcement (p.793) could result in blurring of the line from the law enforcement side, rather from the military. Of course, this line has been blurred in the past during proxy guerrilla wars as part of the cold war, and also in the post-colonial era, however the use of drones means that state can exploit this ‘grey-zone’ (p.794) without the need to co-opt native forces. Further in terms of blurring the line, the articles holistic approach to using all three relevant bodies of law simultaneously (p.795), rather than as discreet and separate bodies could potentially result in a legal blurring as well. This seems to be have happened in relation to drones strikes under Obama, where justifications seem to use language from all three bodies of law.

Also relating to this blurring is the prevention of combatting long-term security threats with long-term military deployments under self-defence justifications (p.801). There is not a problem with this position in law, but the point at which military forces hand over to law enforcement is tricky in practice (UNSC). Handing over to police who may not be able to handle a situation may result in a back-and-forth over the armed conflict threshold, further blurring the line. Also, such a situation would require law enforcement to be able to deal with violent NSAs who may try to retake power or position, and therefore be ‘militarised’ to some degree (as seems to be happening in Afghanistan after the pulling out of most NATO forces). It seems that, whilst perpetual war is of course undesirable for all, the point at which this type of conflict ends may require that blurring for some time longer.


Peter Kalu, Lancaster University

The article posits the triple lock of laws that surround the use of armed drones ( Just War laws /ius ad bellum; international humanitarian law; international human rights law) are sufficiently robust to deal with the changes that use of armed drones brings.

In relation to ius ad bellum, issues of consent and imminent threat, of the idea of war being an exceptional state not a “normal” activity are explored. In particular, the article accepts that armed drones decay the idea of war being an exceptional state.

Regarding IHL, there is an examination of the meaning of protracted armed violence by non-State groups, including tests such as whether inter alia such groups are organised and have a clear command structure. Two highly important areas are scrutinised: (1) the principle of distinction vis a vis determining civilians from combatants. and (2) whether aggregation of  non-state armed groups is permissible  when seeking action against armed groups who are spread over several States. IHRL is also explored, particularly the way it combines with ius ad bellum and IHL as a protective shield of the right to life.

The problem with all such analyses is that such law can be ignored or avoided with impunity by state actors who have the ability to create fake information, to render processes opaque and to evade transparency or fake it. The biggest example of this, indeed the elephant in the room in this area, is the manipulation and avoidance by USA and UK around the ‘weapons of mass destruction’ case for intervention in Iraq. This renders arguments around whether any laws apply similar in their scholastic irrelevance to the ‘how many angels can dance on a pinhead’ debates of the Middle Ages. The authors allude as much when, in examining the question of a State’s the right to attack hostile non-State actors abroad even without the consent of the ‘host’ State, they admit:  ‘the law on this matter appears to have changed’.  It’s worth quoting this in full so the problem can be laid bare:

“Nonetheless State practice relating to the use of force against non-State actors since 2001, is now broad and extensive, and indeed it extends beyond the immediate post 9/11 context. Following this practice, and despite academic controversy, the law on this matter, appears to have changed.”

When to this is added the high degree of blur including in the ‘political imaginary’ that drones bring to so many areas (too great a subject to set out briefly here) the problems compound.

Liam Halewood, University of Liverpool

This article is a great overview of the relevant issues of the use of armed drones, issues that will become more common with the proliferation of armed drones. I am particularly interested by the applicability of the European Convention on Human Rights (ECHR) to targeted killings and would like to add a comment in relation to this aspect of the article (essentially pages 822-825). The article rightly mentions the concerns regarding the applicability of the ECHR to drone strikes:

“one of the key difficulties posed by drones is that the attacking State can engage in targeted killing without exercising effective control over territory or without having the individual in custody, leading to question over whether such persons fall within the ‘jurisdiction’ of the attacking State for the purposes of their human rights treaty obligations”

This statement is a direct reference to the concept of jurisdiction that has developed through the jurisprudence of the European Court of Human Rights (ECtHR) and reflects the concern that targeted killing by drones would not be within the jurisdiction of the ECHR (A viewpoint shared by the UK Government).

The subsequent discussion on how the positive/negative obligations in human rights treaties should apply is interesting but does not represent the law as it stands (see how words such as “should” and “likely” are used. It is an argument rather than a legal position). Essentially, the argument is that the negative obligation to respect the right to life should be territorially unbound and therefore any State killing is within the jurisdiction of the relevant human rights treaty. This is a conception of jurisdiction that I support but it is a legal argument rather than a current legal position. The reference to Milanovic is appropriate because his work on the extraterritorial applicability of human rights treaties is highly regarded but even Milanovic acknowledged recently that drone strikes are “off the table” in regards to being regulated by the ECHR.  For the positive/negative obligation theory to be adopted by the ECtHR, it would take a great expansion of the reach of the ECHR, which the Court has not yet endorsed and has been clearly reluctant to do so.

Nonetheless, it is my view that there is scope to argue that targeted killing by drones would fit within the current conception of jurisdiction under the ECHR. If perceptions of targeted killing by drones change to recognise the technological characteristics of drones to find and track a target (control over an individual), whilst also acknowledging that a targeted killing is a unique counter-terrorism security operation (public powers), then perhaps targeted killings can be regulated by the ECHR, without requiring a drastic expansion by the ECtHR to make it so.

We will be adding further comments from group members as they come in. If you would like to join TTAC21, please e-mail If you don’t want to join the network, but just want to leave a comment about the article, or respond to the thoughts from group members, please do so in the comment box below.

6 thoughts on “The international law framework regulating the use of armed drones – Heyns, Akande, Hill-Cawthorne and Chengeta

  1. Hi Josh,
    Glad you mentioned the special forces angle as this was another one that came to mind when I read this article. And not just special forces, but intelligence agencies such as MI6 / CIA etc. At what point do such operations constitute an act of war? Given what we must assume to be a prevalence of such activities are we at the point where these operations are now a ‘given’ and considered acceptable by the international community? Is there a chance that drones may also go this way? Of course the main difference between drones vs covert operations is the overt nature of the drone, however I do wonder if we will some day reach the point where it will become increasingly difficult to attribute drone strikes to certain actors, especially when rogue states and/or terrorist organisations start to make use of them.


    1. HI Mike,
      The actions of intelligence agencies is a great point. In Gregory McNeals article Targeted Killing and Accountability (, he suggests that Covert action can function as a legal framework for using force by the CIA. Seeing as they do have their own paramilitary, the Special Activities Division, there must be some sort of legal framework for their use. Although as to how formalised it is, I doubt anybody who hasn’t been employed at Langley would know. I don’t think MI6 have the same sort of abilities in-house, but probably link up with UK special forces in various semi-formal ways.

      In terms of an act of war, if a covert operation were launched against a state, an armed conflict would exist from the firing of the first shot. But if launched against a non-state actor, this would only constitute an armed conflict when the ‘violence is sufficiently intense, and the armed groups sufficiently organised’ (para 221, J Kleffner in D Fleck ‘the Handbook of [IHL]’ (3rd edn, 2013)). I think I would agree that the international community probably does begrudgingly just accept that these operations happen, whether by drone, special forces, or deniable forces. But, perhaps they are so useful for all states (or at least the powerful ones) that to try and restrict these activities would just cause political harm and dry-up the back channel communication between states.

      I’m not sure if unattributable drone strikes would happen, after all, if you’re trying to score political or military points, you want everyone to know who did it. But, maybe some groups or individuals would want to take action without any of it leading back to them. If a perpetrator cannot be caught, potential punishments stop being a deterrent to them. Definitely worth thinking about more, particularly the deterrence angle.


  2. At this point I wonder, how important is ‘attributality’ (i.e. the ability to attribute an action to an actor) in regards to international law and more widely, to war? It seems to me that what with the changing nature of global warfare, we have moved to a point now where postmodern war doesn’t even have to be between two defined (State) actors as such, however will always seem to require at least *one*, for otherwise what would be the point (for the terrorists)?

    On a related note, if neither side is ‘attributable’ to an action as such (I’m thinking here for example, covert actions / the intelligence agencies), can such actions ever be an act of war as long as there is mutual ‘wilful blindness’? It would seem to me that this ‘covert reciprocity’ would seem to function to legitimise one’s own similar covert actions…

    [Hope the above makes sense^]


  3. Hi Mike,

    Attributability is generally really important in terms of international law. Considering that the law of armed conflict allows people to kill others and it not be murder, finding out whom can kill and be killed is crucial for legal compliance. When we talk about ‘non-international armed conflicts’, they are between either a state and non-state actor, or between non-state actors. The ICRC do a lot of work trying to convince non-state actors to comply with the law of armed conflict. I think they have had some successes with rebel groups who want to be seen as a legitimate opposition to the state, but groups also willing to engage in terrorism must be a bit harder to convince!

    I think the willful blindness and lack of attributability ideas are fascinating. Sounds like a research project in the making!


  4. Yes, and of course the problem here is that groups engaging in terrorism *don’t* want to me seen as legitimate, so in this case, the ‘carrot’ held out by the ICRC is not very convincing! It strikes me then that in this age of information warfare and global terrorism that the law, and our governing bodies are struggling to content with two completely new ways of thinking.

    (I agree, the wilful blindness thing is very interesting!)


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s