Scope of the prohibition

The problem of human shields in an international armed conflict is addressed in a number of provisions of the Geneva Conventions.Footnote ¹³ The ban on using civilians for this purpose was picked up and extended in 1977 by Article 51(7) of Protocol I:

Whereas in 1949 the ban was limited to the scope of the Third and Fourth Geneva Conventions and therefore concerned only prisoners of war and ‘protected persons’, Article 51(7) of Protocol I concerns and consequently protects the civilian population as a whole. The scope of the ban on human shields is thereby extended or clarified not only ratione personae but also ratione materiae. The ‘presence’ of civilians being used as human shields covers two types of situation: those where civilians are placed on or close to military objectives and those where military objectives are placed in the midst of civilians. Article 51(7) also covers cases where ‘movements’Footnote ¹⁴ of the civilian population are used to cover military operations.Footnote ¹⁵

Finally, in the 1998 Rome Statute of the International Criminal Court, the use of human shields during an international armed conflict is classified as a war crime (Article 8(2)(b)(xxiii)).

To demonstrate the existence of a constant and uniform practice and an opinio juris within the international community, one possible source is the jurisprudence from the war crimes tribunals that followed the Second World War.Footnote ¹⁶ A more recent source is the case of Prosecutor v. Radovan Karadzic and Ratko Mladic before the International Criminal Tribunal for the former YugoslaviaFootnote ¹⁷ (ICTY). One may also refer to the military manuals of certain states, which outlaw the practice, as well as the domestic law of some countries and to a number of official statements by states, and by the ICRC, condemning it, which carry a certain weight in the absence of any contrary practice.Footnote ¹⁸

There is no treaty-based rule that expressly prohibits the use of human shields in non-international armed conflicts. Article 5(2)(c) of Additional Protocol II simply provides, as does Article 19 of the Third Geneva Convention in the case of international armed conflict, for the evacuation of persons deprived of their liberty from combat areas so that they are not exposed to danger. In addition to that provision, the ban on using human shields could be covered by the scope of Article 13(1) of Protocol II, which stipulates that the civilian population and individual civilians must enjoy ‘general protection against the dangers arising from military operations’. However, it is customary law that provides the best basis for asserting that the use of human shields is also prohibited in non-international armed conflicts. The prohibition follows from the fundamental obligations to distinguish between combatants and civilians and to take the relevant precautions. Moreover, the use of human shields is often placed on the same footing as hostage-taking, which is forbidden by customary law and also by Article 4(2)(c) of Protocol II.Footnote ¹⁹ Under the Rome Statute, use of human shields in an international armed conflict can be prosecuted as a war crime. However, this is not the case where the act is committed in a non-international armed conflict.

There is no provision of international human rights law that expressly forbids the use of human shields outside situations of armed conflict. However, it seems logical that such prohibition would fall within the scope of the core fundamental rights such as the right to life or the prohibition of torture and other cruel, inhuman or degrading treatment.Footnote ²⁰ The physical and mental suffering inflicted, and particularly the fact that such a person is awaiting near-certain and imminent death,Footnote ²¹ can also serve as a basis for this assertion. Moreover, although the accused in question were being tried for war crimes, the ICTY has also expressed the view on several occasions that the use of human shields constituted ‘inhumane treatment’ and ‘cruel treatment’.

Recurring questions

As we have seen, although the prohibition on the use of human shields is plainly set out, it is nevertheless frequently breached. The question as to how such breaches are to be categorized has arisen on a number of occasions. Moreover, although the prohibition is clear, some of the legal problems associated with it, particularly those concerning the phenomenon of voluntary human shields, appear to be rather less so.

Using human shields seems to be a very ‘treacherous’ practice, but does it qualify as an act of perfidy in legal terms? Some experts hold that, in unbalanced armed conflicts, the weaker party is naturally prompted to hide in more densely populated areas so as to make it more difficult or impossible for the adverse party to identify military objectives.Footnote ²² However, ‘if the civilian population is intentionally used as a kind of shield to protect defending units, there is no longer any question of permitted deception: this clearly amounts to prohibited perfidy’.Footnote ²³ The commentaries to Article 28 of the Fourth Geneva Convention could also be interpreted as asserting the same thing, as they contain a lengthy passage distinguishing the use of protected persons as human shields from a ruse of war.Footnote ²⁴ Yet, traditionally, ruses of war are always opposed to acts of perfidy.

In the present author's view, despite the absolute prohibition on the use of human shields, to apply the term ‘act of perfidy’ to such acts is not legally appropriate, considering the definition given to the term ‘perfidy’ in international humanitarian law. This definition contains three elements. Two of them are subjective, namely an appeal to the opponent's good faith and the intention to deceive that good faith. The third, objective, element is that the deceit must concern the existence of a protection granted by international humanitarian law.Footnote ²⁵ The use of human shields does indeed rely on the protection accorded to civilians or prisoners of war by international humanitarian law, but the enemy is not deceived. The protected status of the people so used is not feigned but genuine. Moreover, for an act to fall within the definition of perfidy as prohibited by Article 37 of Protocol I, it must have been committed with the intention of killing, wounding or capturing an enemy.Footnote ²⁶ Some authors go even further, stating that ‘if it is used solely for combat against military objects, for example, without affecting any enemy combatant, it is permissible’.Footnote ²⁷ In the case of human shields, the aim is not to kill, wound or capture the enemy, but to defend military objectives against attack. It would therefore appear that the use of human shields does not fit the definition of an act of perfidy and still less that of an act of perfidy as defined in Article 37 of Protocol I and the customary rule codified therein.

The problem of voluntary human shields was not explicitly addressed by the drafters of the Conventions and their Additional Protocols. Does that mean that voluntary human shields are not covered by the prohibition on human shields? That seems unlikely. The party benefiting from their presence does not necessarily need to have gone looking for them and stationed them forcibly in front of a military target. Nor is it required that the people used as human shields should be unaware of the fact.Footnote ²⁸ On the contrary, it appears that this prohibition ‘applies both when the civilians are hostages and when they have volunteered to shield military targets’.Footnote ²⁹ The essential element in the prohibition on use of human shields is rather the intention to use the presence of humans as shields to shelter a military objective. This is corroborated by the International Criminal Court's Elements of Crimes.Footnote ³⁰ The material element there is moving or taking advantage of the location of protected persons; the mental element is the intention to shield a military objective from attack or to shield, favour or impede military operations. There is no trace of any requirement that there be ignorance on the part of the people concerned or that they be constrained. However, the criterion of intention is always tricky. How can one be certain that tolerance or inaction in relation to people who have voluntarily stationed themselves on strategic sites should be understood as an intention to take advantage of their presence to shield a military objective from attack? However, a finding of intention can often be arrived at inductively on the basis of the factual circumstances.Footnote ³¹

If the authorities allow such a thing to happen without taking any action, that could be considered all the more revelatory of an intention to use human shields, since, in addition to the absolute negative obligation never to do so, the authorities also have positive obligations, albeit relative ones this time, to take various precautionary measures, including keeping civilians away from military targets.

The difficulty to consider acting as a voluntary human shield as tantamount to taking direct part in hostilities

Experts are very divided on the question of whether or not acting as a voluntary human shield is tantamount to taking direct part in hostilities.Footnote ⁴⁷ Nevertheless, there are a number of jurists who do take the view that acting as a voluntary human shield – that is, deliberately trying to ward off an attack on a military objective – is indeed tantamount to taking direct part in hostilities.Footnote ⁴⁸ It follows from this that voluntary human shields lose their protection as civilians because they are taking direct part in hostilities.Footnote ⁴⁹ According to this analysis, human shields are acting in exactly the same way as anti-aircraft defence systems, only more effectively.Footnote ⁵⁰ The High Court of Justice of Israel took a similar position in a recent judgment on targeted killings of ‘terrorists’, in which it examined the question of what the law is regarding civilians serving as a ‘human shield’. The Court found that ‘if they do so of their own free will, out of support for the terrorist organization, they should be seen as persons taking direct part in the hostilities’.Footnote ⁵¹ Others do not share this view, claiming that it would be incorrect to state that people who place themselves voluntarily in front of a legitimate target are taking direct part in hostilities.Footnote ⁵²

Hostile acts do not necessarily involve the use of weapons and taking direct part in hostilities includes ‘attacks’.Footnote ⁵³ Attacks include offensive and defensive acts.Footnote ⁵⁴ It would therefore be possible to consider that voluntary human shields who place themselves unarmed in front of military objectives in order to ward off an attack, in other words to defend it, are taking direct part in hostilities.

However, ‘hostile acts’ and taking ‘direct part’ in hostilities can be defined as ‘acts which by their nature and purpose are intended to cause actual harm to the personnel and equipment of the armed forces’ of the enemy.Footnote ⁵⁵ If we adopt this definition of taking direct part in hostilities, the actions of the voluntary human shield do not fit it very well. The voluntary human shield does not strike the enemy forces; he merely protects by a passive attitude the personnel or hardware of his own armed forces. One could extrapolate and say that he strikes the enemy forces indirectly, since by protecting his own military hardware and personnel, he preserves them and enables them to strike the enemy. However, this interpretation does not square with the use of the word ‘actual’ and still less with the Commentaries, which describe direct participation in hostilities as acts that ‘present an immediate threat to the [adverse] party’.Footnote ⁵⁶ Even if it could be said that human shields present a threat to the adverse party because they protect military targets that themselves present such a threat, this threat is not immediate. Although it could certainly contribute to a state's ‘war capabilities’, the human shield's participation is only indirect. The voluntary human shield cannot therefore be regarded as taking direct part in hostilities.

However, the degree of actualness or immediacy a threat must possess in order to count as direct part in hostilities can be difficult to define. The High Court of Justice of IsraelFootnote ⁵⁷ has taken the view that ‘the “direct” character of the part taken should not be narrowed merely to the person committing the physical act of attack. Those who have sent him, as well, take “direct part”. The same goes for the person who decided upon the act, and the person who planned it’. This notwithstanding, ‘[d]irect participation in hostilities implies a direct causal relationship between the activity engaged in and the harm done to the enemy at the time and the place where the activity takes place.’Footnote ⁵⁸ The essential difficulty, however, is to determine the intensity required to create this causal relationship. It would seem important not to take too broad an approach: to interpret it too loosely would lead to voluntary human shields easily being placed on the same footing as people taking direct part in hostilities, which would mean, as some experts have pointed out, that they could be attacked ‘during their preparation, namely when moving towards the military objective to be shielded by their presence’.Footnote ⁵⁹ There is, however, no military necessity to attack them when they are not protecting a military target. Nor is there any point in targeting the human shield himself in addition to the military target he is trying to protect!Footnote ⁶⁰ If the causal relationship is stretched too far, just about anything could be seen as taking direct part in hostilities, including the attitude of the general civilian population, because by undermining the morale of the population, it is possible to weaken its allegiance to the state, and hence the state itself. In that case the fundamental principle of distinction would disappear and we would be giving carte blanche to total war. Finally, as some experts point out, the causal relationship should not be read too broadly because the law must be applicable on the ground.Footnote ⁶¹

At all events, we can safely say that the acts of voluntary human shields do not square easily with what we can consider as taking direct part in hostilities. Moreover, to argue the contrary seems utterly self-defeating. The whole point of human shields is that their civilian status, that which makes them ‘unattackable’, is used to protect a military objective. However, if they were taking direct part in hostilities, they could themselves be an object of attack. Their presence in front of a military target would therefore be entirely pointless. In that case, no one would voluntarily position himself in front of a potential target! At the same time, one can have the feeling that, from the point of view of direct participation in hostilities, there are voluntary human shields and voluntary human shields. There seems to be little common ground between a human shield on the roof of a missile launching pad and, to take an example cited by an expert, ‘a Somali woman who walked across the street holding her arms up and hiding behind her one or two fighters, who would fire their weapons from behind her flowing white gown’.Footnote ⁶² The ICTY in the Tadić case adopted a case-by-case approach:

We cannot therefore affirm in general and absolute terms that a voluntary human shield is or is not taking direct part in hostilities. This can only be ascertained by an appraisal in concreto of the way in which the human shield indeed tries to protect the military objective in question.

When it comes to determining whether a voluntary human shield, in defending a military objective, is or is not taking part in hostilities, the criterion of the real, physical shield he constitutes against a potential attack seems to be an interesting avenue of approach. For example, some minimum fighting would be required from an infantry division to get through a crowd of people standing between it and a military objective. In contrast, a small number of human shields standing near a military objective to protect it from an air strike do not constitute a real obstacle for the attacking party in the material sense of the word. In this case, the presence of a human shield is only a legal obstacle for the attacker, who hesitates to attack only out of fear of violating international humanitarian law with the attendant political and media impact. It would therefore seem overbold to declare an obstacle of that nature to be direct participation in hostilities.

The inherent risk run by the voluntary human shield

As a voluntary human shield cannot in most cases be regarded as taking direct part in hostilities, he retains the full protection due to a civilian status. It is plain, however, that his position in front of a military objective makes his situation more dangerous than that of a civilian who is nowhere near any potential military targets. His condition can be compared with that of workers working in armaments or munitions factories.Footnote ⁶⁴ As civilians who take only an indirect part in the hostilities, these workers continue to enjoy protection.Footnote ⁶⁵

The idea of the ‘ordinary’ risk to which these civilians are exposing themselves is also to be found in the writings of Frédéric de Mulinen: ‘Les personnes civiles se trouvant à l'intérieur d'un tel objectif [d'un objectif militaire] ou dans son environnement immédiat partagent le danger auquel il est exposé.’Footnote ⁶⁶

In the specific case of voluntary human shields, Josiane Haas seems to be of the same opinion:

She goes on to say that the attacker is, of course, always bound by the principles of distinction and proportionality. In short, human shields do not lose their protection as civilians. They are simply close to a military objective and, as a result, ‘these civilians will bear the risk of falling victim to a legitimate attack on the shielded object’.Footnote ⁶⁸

Is the voluntary human shield a war criminal?

As we have seen, using human shields is one of the acts classified as a war crime by the Rome Statute of the International Criminal Court (Art. 8(2)(b)(xxiii)). We may therefore ask ourselves whether a person who places himself in front of a military objective is also committing a war crime, as that person is after all misusing his status as a civilian.

According to Elements of Crimes, adopted in 2002, the elements of the war crime of using protected persons as shields are:

1. 1. The perpetrator moved or otherwise took advantage of the location of one or more civilians or other persons protected under the international law of armed conflict.

2. 2. The perpetrator intended to shield a military objective from attack or shield, favour or impede military operations.

3. 3. The conduct took place in the context of and was associated with an international armed conflict.

4. 4. The perpetrator was aware of factual circumstances that established the existence of an armed conflict.

The three last conditions can perfectly well be met by a voluntary human shield. However, the first element poses a problem. It is fairly clear from the wording that the drafters were referring to people who had moved or taken advantage of the presence of protected persons to shield a military objective, and, in the case of voluntary human shields, not the shields themselves. The elements of this crime suggest that the prohibition refers to taking advantage of other civilians or exposing them to danger. It should be borne in mind that the principles of criminal law are to be interpreted restrictively and that reasoning by analogy is not allowed. Moreover, if a civilian taking direct part in hostilities is not ipso facto a war criminal, we may ask ourselves a fortiori why a human shield who, as we have just seen, can scarcely be regarded as taking direct part in hostilities, should be regarded as such.Footnote ⁶⁹

The case of involuntary human shields

If the concrete and direct military advantage anticipated proves sufficiently significant in relation to the potential damage to the human shields, the attack may take place. The appraisal must be in concreto and must always take into account the military advantage, which can shift, for example over time (a bridge of fundamental importance one day may have little strategic significance the next day depending on where the front line lies at the time), and the extent of collateral damage that will be caused to civilians. The extent of that damage may depend, among other things, on the number of civilians likely to be affected. A particular military advantage may thus be regarded as sufficiently important to justify an attack on a military objective protected by five people being used as human shields. However, if the object is protected by five hundred human shields, the outcome of the deliberation may change. In practice then, there is no doubt that in certain cases, the attacked party can effectively protect the military objective by placing sufficient human shields on or around it! Some authors understandably reject this conclusion, which they consider shocking, and recommend human shields not be taken into account by a commander applying the proportionality principle while preparing for or deciding on an attack.Footnote ⁷⁷ For example, the commander could apply the proportionality test in such a way as to comply a priori with Articles 57(2)(a)(iii) and 51(8) of Protocol I, but the human shields would simply not be taken into account on the side of the equation concerning damage to civilians. However, I do not consider this interpretation to be correct. Civilians, whether they are human shields or not, cannot simply be left out of the equation. The fact that civilians are close to a military objective because the attacker has breached his obligations makes no difference. It would be against both the spirit and the letter of Article 51(8) if civilians were to ‘pay’ for the wrongs of a belligerent party.Footnote ⁷⁸

The case of voluntary human shields

According to Michael N. Schmitt, since voluntary human shields are to be regarded as taking direct part in hostilities, they lose their protection as civilians and can themselves be an object of attack. Accordingly, ‘voluntary shields … are excluded in the estimation of incidental injury when assessing proportionality’.Footnote ⁷⁹ However, the same author remarks that there is no military necessity to attack the human shield per se. The necessity is attached only to the military objective he is protecting. Accordingly, ‘the only practical impact of their willingness to serve as shields is that they need not be included in proportionality calculations’.Footnote ⁸⁰ As we have already seen, I do not think that a voluntary human shield should be regarded as taking direct part in hostilities. Schmitt's statement nevertheless raises an interesting question, namely whether the voluntary nature of their presence can have an impact on the appraisal of proportionality. Intuitively, one could indeed have the feeling that a human shield who is in front of a military objective by his own choice should not count in the same way as someone chained to a military objective against his will. However, if we were to take into account the voluntary nature of a human shield's action, it would be tantamount to assigning a lesser value to voluntary human shields. In mathematical terms, in an equation that balances military advantage and damage to civilians, a lower coefficient would be applied to the weighting of the civilian side of the equation. An approach based on the human shield's willingness or otherwise would also run the risk of dangerously eroding civilian protection. If a human shield should have less weight in the equation because he has chosen to guard a military objective, a civilian who remains close to a military objective despite warnings should also weigh more lightly in the balance. The same would apply to a civilian who lives close to an obvious potential military objective such as a military barracks but who has chosen not to leave his home. There would be a real risk of this approach being abused by the attacker, who could be tempted to classify all civilians close to a military objective he is targeting as voluntary human shields, if as a result he could make his proportionality appraisal on a more flexible basis.

Furthermore, could not we consider that this method of taking a civilian's willingness into account should be regarded as contrary to the principle of inalienability of rights, enshrined in Article 8 of the Fourth Geneva Convention, for example, which states that protected persons ‘may never renounce … the rights secured to them’ by that Convention. Would this not include the principle that subtends those rights, namely the fact that civilians may not be attacked unless they are taking direct part in hostilities? To conclude on this point, I would refer to another practical element that pleads in favour of treating human shields the same way, whether they are voluntary or not. For the commander faced with the reality on the ground, it is not always easy to distinguish civilians placed in front of a military objective against their will from those who are standing there of their own volition. Except in cases where the voluntary nature of the human shield's action has been given media coverage, or at least brought to the attacker's attention, how is he to distinguish between a willing human shield and an unwilling one? It is very difficult to determine whether the human shield is really in front of the military objective of his own volition. An individual who appears to be acting of his own free will may turn out to have been simply giving in to pressure.Footnote ⁸¹ In short, although a distinction based on willingness could have some relevance in a criminal case, it has no place in the conduct of hostilities as it cannot be applied on the ground.

The obligation to take constant care to spare civilians when conducting military operationsFootnote ⁸³

Article 57(1) has a broader scope than the subsequent sections, which specifically concern ‘attacks’; paragraph 1 applies to ‘military operations’ in a more general way. The term covers ‘any movements, manoeuvres and other activities whatsoever carried out by the armed forces with a view to combat’.Footnote ⁸⁴ Over and above the specific precautions set out in the subsequent paragraphs of Article 57, the parties must therefore also take constant care to spare human shields.

The obligation to verify that the objectives to be attacked are military objectivesFootnote ⁸⁵

In the case of human shields, we have seen that a sufficiently significant military advantage in relation to the danger to which human shields are exposed could render an attack on a military objective legitimate despite their presence. It is therefore all the more vital to be sure of the military nature of the objective, as attacks on civilians and civilian property are categorically prohibited. The information to be gathered in the course of this verification concerns not only the nature of the target itself but also its environment. As we have seen, even in the presence of a military objective, an attack can prove to be prohibited, for example if far too many civilians are being used as human shields and would be endangered by the attack in relation to the size of the military advantage to be derived from it. One particular difficulty is raised by ‘emerging targets’.Footnote ⁸⁶ In contrast with planned operations, an ‘emerging target’ situation calls for an instant determination of the military nature of the target and the conduct to be adopted if it is protected by human shields. The commander is required to ‘do everything feasible’ to verify the nature of the objective, as no one can be obliged to do the impossible.

The obligation to choose means and methods of attack with a view to avoiding, and in any event minimizing, incidental loss of civilian life, injury to civilians and damage to civilian objectsFootnote ⁸⁷

For example, to comply with this obligation, and still remain subject to ‘feasibility’, if human shields are not in front of the military objective all the time, an attack should be launched at a time when they are not present. Similarly, where the military objective is protected by human shields, the attacker should use weapons that will destroy the target without harming the human shields around it or will harm them as little as possible. This means that the attacker should try to avoid using ‘missiles and remotely controlled weapons … (unless the attack uses the new generation of remotely controlled “precision-guided munitions”), since the targeting capabilities of remotely controlled weapons are traditionally extremely bad’.Footnote ⁸⁸ Although this does not mean that the parties have to acquire precision-guided weapons, even if they have the means to do so, if they do have such weapons certain situations require that they use them where it is possible and feasible to do so.Footnote ⁸⁹

The obligation to cancel or suspend an attack if it becomes apparent that the objective is not a military one or is subject to special protection or that the attack would be disproportionateFootnote ⁹⁰

Unlike Article 57(2)(a), this obligation is addressed not only to those planning or deciding on an attack but also to those carrying it out. It may be difficult for someone carrying out an attack to assess its proportionality, as the part he is carrying out may seem disproportionate while the attack as a whole remains fully in compliance with the proportionality principle. However, when it comes to the obligation set out in Article 57(2)(b), the role of those carrying out the attack remains essential, as there may have been a mistake or fresh information may appear and change the original premises. For example, if a pilot has received the order to bomb an objective but he realizes at the last minute that it is protected by a human shield, he should suspend the attack and refer back to his command.

The obligation to give effective advance warning of attacks which could affect the civilian population ‘unless circumstances do not so permit’Footnote ⁹¹

Can this requirement be regarded as met when the attacker issues a general list of the various items of infrastructure he considers to be military objectives?Footnote ⁹² Although it might seem possible to reply in the affirmative, there will be cases when efforts should be made to give a more specific warning. When the military objective is protected by human shields, a warning before an attack on the objective will let the party using the human shields – thinking that they will forestall an attack – know that the stratagem has not worked, and give it a chance to remove the human shields from the target. Voluntary human shields might also discover in this way that their presence will not prevent an attack on the military objective they are guarding and have the time to leave the area. ‘Effective advance’ warning must be given that is sufficiently in advance to allow the evacuation of civilians, including human shields. However, care should be taken not to give the warning too far in advance. Otherwise civilians may think that the danger has passed, and human shields, or those who have positioned them, may think that their presence has forestalled the attack, whereas the attack will still happen but a little later. Finally, it should be remembered that complying with the obligation to warn does not release the attacker from his duty to take other precautionary measures. He cannot therefore consider an entire area as a military objective simply because he has recommended that it be evacuated. Even if he were to regard any civilians remaining as voluntary human shields, they would, as argued above, still enjoy the same protection as any other civilians.

The obligation, when there is a choice between two military objectives for obtaining a similar military advantage, to choose the one which may be expected to cause the least danger to civiliansFootnote ⁹³

This rule, referred to as the choice of ‘the lesser evil’,Footnote ⁹⁴ leaves much to subjective appreciation, particularly as the two targets are legitimate military objectives.Footnote ⁹⁵ A choice of this kind is generally only possible where the military objectives in question are lines of communication. For example, if human shields are positioned on a bridge and the communication line can be broken by attacking another bridge that is not surrounded by civilians, the obligation set out in Article 57(3) obliges the attacker to take that option.