Mittwoch, 18. Mai 2016

NATO Paper on the TALLIN MANUAL Paper 5 The Nature of International Law Cyber Norms summary english


The Nature of International Law Cyber Norms

Michael N Schmitt and Liis Vihul

Executive Summary

The Nature and Place of International Legal Norms

International law is typically described as prohibitory in nature: any activity that is not disallowed is generally permitted. But even when law does exist, it may prove lacking when meeting unanticipated circumstances and thus is occasionally breached as part of the process of creating a new norm. Indeed, it is often said that customary law norms are “made in the breach”. By way of illustration, it may be that pre-existing human rights law would, if logically applied in the cyber context, prohibit intrusions into certain forms of cyber communications between individuals. However, if states treat this customary norm as inconsistent with their need to ensure, for instance, the security of their cyber systems, they may begin to act contrary to the norm. Over time, their state practice, could, as will be explained, be viewed by states as legal, such that the original human rights norm will have been modified. Given the novelty of cyber activities, they are particularly vulnerable to this dynamic of customary law. Once the international law boundaries of conduct are demarcated, domestic legal, political (policy), ethical and other norms can operate to further restrict or require particular conduct in cyberspace. For instance, while it is unclear precisely how international human rights norms in the realm of privacy restrict state monitoring of personal cyber communications, monitoring may constitute a violation of domestic constitutional law or be contrary to state policy or the ethical benchmarks that a state has adopted. Thus, international legal norms merely define the space within which states may engage in normative construction. Of course, states may act to transform these non-legal norms into those with legal authority by adopting a treaty incorporating them or engaging in state practice that crystallises over time, as described below, into customary law.

Terminological Precision

It is clear that when cyber operations accompany kinetic hostilities qualifying as armed conflict (as with the conflict between Russia and Georgia in 2008 or that taking place in Syria at the time of writing), IHL applies fully to all the cyber operations that have a nexus to the conflict, whether they are launched by states, non-state groups or individual hackers. For instance, in the same way that IHL prohibits injurious or destructive kinetic attacks against civilians and civilian objects, it likewise prohibits cyber attacks against them having the same effects.17

For international lawyers the term “cyber war” is better rendered as “cyber armed conflict”. When non-lawyers speak of the norms applicable in cyber war, the lawyer will accordingly insist on examining the attendant circumstances, because only if they qualify as armed conflict will the specific international law norms applicable therein attach. Otherwise, the situation will be subject to those aspects of international law that apply during peacetime, such as the law of state responsibility and human rights law.

The second term that causes confusion between the normative communities is, again, “attack”. As noted, “armed attack” is a legal term of art in the jus ad bellum. Yet, “attack” is also a legal term of art in IHL. The term does not simply refer to military operations directed by one belligerent against another during an armed conflict. Rather, it is defined in Article 49 of Additional Protocol I to the Geneva Conventions as “acts of violence against the adversary, whether in offence or in defence.”18 The Tallinn Manual accordingly defines a cyber attack as “a cyber operation, whether offensive or defensive, that is reasonably expected to cause injury or death to persons or damage or destruction to objects.”19

The definition of an “attack” lies at the core of IHL, because many of its prohibitions are framed in terms of prohibition of attacks, the paradigmatic examples being those on directing attacks against civilians and civilian objects.20 To the extent that a cyber operation does not qualify as an attack in the IHL regime, the prohibitions are inapplicable. Consequently, when a non-lawyer uses the term “cyber attack”, clarification must be sought (in addition to the jus ad bellum issue outlined above) not only as to whether the operation occurred during an armed conflict such that IHL applies, but also whether the operation constitutes an attack such that IHL prohibitions and restrictions come into play.

Clearly, terminological indistinctness and imprecision have long hobbled interdisciplinary understanding between the legal and non-legal communities; they continue to do so today. A proper grasp of the international law governing cyber operations, and its likely future evolution, demands terminological fastidiousness. It is to that law that we now turn.

General Rules Governing Treaty Law

 

In addition to reservations, states may issue interpretative declarations that clarify their position with regard to a particular provision of the treaty or to how the treaty will be applied by the states concerned. Declarations have no technical legal effect on the state’s rights or obligations. However, states sometimes make interpretative declarations that de facto amount to reservations. For example, the United Kingdom has issued a statement concerning the prohibitions on reprisals set forth in Additional Protocol I to the 1949 Geneva Conventions.33 The declaration arguably denudes certain provisions of their effect. Thus, declarations, like reservations, must always be carefully surveyed when evaluating the actual normative reach of a treaty.

Perhaps the most important aspect of treaty law deals with interpretation, as a treaty’s text may be vague or ambiguous. Such ambiguity is often the only way the parties involved were able to achieve sufficient consensus to adopt the instrument. The Vienna Convention on the Law of Treaties provides that treaties “shall be interpreted in good faith in accordance with the ordinary meaning to be given to the terms of the treaty in their context and in light of its object and purpose.”34 The term “context” refers to the other text of the treaty, as well as to any agreement between the parties made at the conclusion of the treaty.35 In addition to context, interpretation of a treaty’s provision should take account of any subsequent express agreement between parties as to its meaning, as well as “subsequent practice in its application that establishes the agreement of the parties regarding its interpretation.”36 If the meaning of a provision remains ambiguous, reference may be made to the “preparatory work of the treaty and the circumstances of its conclusion.”37 In other words, it is appropriate to explore what was in the mind of the parties at the time when the agreement was negotiated and adopted.

Treaty Law in the Cyber Context

Given that cyber activities are relatively new, very few treaties deal directly with them. Prominent contemporary examples include the Convention on Cybercrime,38 its 2006 Additional Protocol,39 the Shanghai Cooperation Organisation’s International Information Security Agreement,40 and the ITU Constitution and Convention41 and International Telecommunication Regulations.42 The rules regarding treaties apply fully to each of these instruments and others that exist or are to be adopted in the future. Since it is not the purpose here to examine their substantive content, it suffices to recall that when considering the formation, interpretation and application of cyber treaty norms, the key guidance is to be found in the Vienna Convention on the Law of Treaties and in the customary law of treaties.

In light of the paucity of cyber-specific treaties, the threshold question is, of course, whether non-cyber-specific instruments even apply to cyber activities. A number of states, including Russia and China, have previously expressed some reluctance to acknowledge that existing international agreements extend to cyberspace.43 This disinclination seems to have been partially overcome in Considering the broad acceptance of the premise that non-cyber-specific treaty law can apply to cyberspace, an array of international agreements that govern state activities in general also constrain cyber activities. As an example, the 1982 Law of the Sea Convention delineates the type of activities that the vessels of one state may engage in while in the territorial sea of another state.48 Although the vessels have a right of passage though the territorial sea, the passage must be “innocent”, that is, not be contrary to the interests of the coastal nation. Conducting cyber operations against the coastal state from aboard naval vessels would consequently violate the innocent passage regime for states party to the Convention, even though that treaty was adopted well before the advent of sea-based cyber operations. Similarly, the 1963 Moon Treaty provides that the Moon and other celestial bodies are to be used for “exclusively peaceful purposes”.49 Therefore, military cyber operations may not be launched from the moon or other celestial bodies, again despite the fact that the treaty predates the technical capability to do so. In Europe, the 1950 European Convention on Human Rights (in effect since 1953) is playing a prominent role in privacy and data protection debates involving cyber communications that its drafters could not have envisaged. well as its shortcomings, in the cyber context. The first deals with the meaning of the term “use of force” in the UN Charter’s Article 2(4) prohibition thereof. The object and purpose of the provision was self-evidently to limit the circumstances in which states might resort to force to resolve their differences. All of the Tallinn Manual experts agreed that a cyber operation by one state against another that causes injury or death to individuals, or damage or destruction to property, qualifies as a use of force. However, no consensus could be reached on the exact threshold at which a cyber activity crosses into the use of force. The International Group of Experts could only offer indicative factors that states are likely to consider when deciding how to legally characterise a cyber operation in this respect.53 Delineations of factors should prove useful as states estimate how their activities will be seen by other states, as well as when they assess the actions of other states against the norm, but they are not legal criteria per se. The object and purpose of Article 2(4) provided a guide to interpretation in the cyber context, but not a fully comprehensive one.

Second, Article 51 of the UN Charter provides that states may use force in response to an “armed attack”. Here, the object and purpose was to ensure that states did not remain normatively defenceless should the enforcement regime established in the Charter fail to operate as planned. But the interpretation of this article remains a source of some uncertainty and controversy because it is unclear whether the right of self-defence extends to attacks conducted by non-state actors, or whether states are limited to law enforcement measures in responding to such hostile acts. This is an issue that was brought to the forefront of international law debate in the aftermath of the 9/11 attacks against the United States by al Qaeda. It is a central one with respect to cyberspace, because a non-state group’s or individual’s capability to launch a hostile cyber operation at a state at the armed attack level is much more likely in the cyber context than the kinetic, due to the relative ease of acquiring the expertise and equipment for a cyber armed attack compared to a kinetic one.54

Recently, both the United States and the Netherlands have taken the position that defensive use of force in the cyber context is permissible under Article 51 even if a cyber-attack by a non-state actor cannot be attributed to another state.55 Those states and commentators who take the more restrictive approach in applying Article 51 to terrorist strikes would likely be at least as restrictive when considering cyber operations mounted by non-state actors. This illustrates that difficulties in interpreting treaty law in the non-cyber context are highly likely to resurface in the cyber context.

It is also unclear when a cyber operation is severe enough to be regarded as an armed attack in the sense of Article 51. According to the Tallinn Manual, operations causing significant damage, destruction, injury or death do qualify. Inclusion of such consequences is consistent with the UN Charter’s object and purpose of limiting the use of force in international relations, but consensus among the International Group of Experts stopped there; the group could not agree on any “bright line test” for determining when such harm is sufficiently “grave” to cross the armed attack threshold.56 Some experts took the position that the term should include operations that cause severe non-physical harm, such as cyber operations directed at crippling a state’s economy.57 Others resisted such a broad interpretation on the grounds that it ran counter to the Charter’s presumption in favour of non-forceful resolution of international disputes. Again, a reliable interpretation of a treaty provision in the cyber context proved elusive because multiple reasonable interpretations were possible.

The third and fourth examples derive from IHL. The paradigmatic interpretive hurdle in IHL is that cited above, the meaning of the word “attack”, which is found in various prohibitions set forth in Additional Protocol I. For instance, pursuant to express provisions of that treaty, it is unlawful to attack civilians, civilian objects, and certain other protected persons and objects.58 Additionally, states are required to consider expected collateral damage at the attack level when assessing the proportionality of their operations,59 and must take precautions to minimise such damage whenever they conduct attacks.60 Interpretation of the term “attack” in the cyber context is essential because, to the extent to which a cyber operation fails to qualify as an attack, these and related IHL provisions do not apply.

Recall the Article 49 of Additional Protocol I definition of attack as an act of violence and the definition of cyber attack found in the Tallinn Manual as “a cyber operation, whether offensive or defensive, that is reasonably expected to cause injury or death to persons or damage or destruction to objects.” All members of the International Group of Experts agreed that Additional Protocol I’s provisions referring to attacks included such cyber operations because they were violent in the sense of Article 49. However, members of the group differed on whether, and if so how far, the notion of violence should be stretched to include operations having non-kinetic effects. Some experts were of the view that the notion is strictly limited to cyber operations that cause physical damage or injury; other operations were not violent and therefore did not qualify as attacks. But a majority of them looked to the object and purpose of the Protocol and its relevant provisions to interpret the term more liberally as applying to a situation in which the functionality of an object is affected by a cyber operation without physical damage having occurred. Illustrating the difficulties that attend the application of treaty provisions to situations that were not envisaged by the drafters, there were differences of opinion within the majority as to how “functionality” should be interpreted.61 As this example illustrates, layers of interpretation can exist.

Finally, a similar IHL-based debate is underway as to whether the term “civilian object” extends to data.62 If so interpreted, a cyber operation designed to destroy civilian data would be prohibited by Article 52 of Additional Protocol I, which bans direct attacks against civilian objects. If not, civilian data is a lawful object of attack, except in those circumstances where its loss might cause physical damage to objects or injury to persons. The critical and unresolved fault line in the debate lies between interpretations that limit the term to entities that are tangible, which is arguably the plain meaning of the term “object”, and those based on the argument that in contemporary understanding the ordinary meaning of “object” includes data.63

These examples illustrate that even strict application of the rules of treaty interpretation set out above fails to fully suffice in adding the requisite clarity when extant treaty provisions are applied to cyber activities. Such interpretive dilemmas are only likely to be resolved over time. Interpretive clarity will be fostered through the recurrent practice of states in application of the provisions in question, including when those states are acting in their capacity as members of international organisations like the United Nations, European Union and NATO. Also relevant will be state expressions of opinion as to proper interpretation of the terms and provisions in question. Recent examples include those proffered by former US Department of State legal adviser Harold Koh64 and by the Dutch Government in response to the AIV report, both of which set forth state positions on the meaning of key aspects of relevant treaty law.65 Judicial interpretation could potentially also shape the meaning of uncertain treaty norms in the cyber context, much as the judgments of the International Second, in the early days of a new technology, states will be reluctant to bind themselves to particular rules until they fully understand how those rules may play out as the technology continues to develop. In particular, there is presently little support for proactively addressing cyber weaponry and cyber military operations. As with all other methods and means of warfare, states are hesitant to restrict the use of weapons that may afford them an advantage on the battlefield until they have sufficient experience to allow them to weigh the costs and benefits of prohibitions and limitations on their use.68

Third, to the extent that states wield cyber capabilities that are strategically or operationally useful, they have an incentive to retain the option of employing them. But those same states may be vulnerable to hostile operations by other states using similar capabilities. Therefore, it may be difficult for a state’s political and legal organs to agree on how the state should characterise a particular practice, as they may view the state’s national interests from different perspectives.

A fourth factor rendering cyber treaties unlikely in the near term is the difficulty of verifying compliance with their terms and effectively enforcing them. To begin with, it is sometimes difficult to even ascertain that harm is the result of a cyber operation. Not only are the technical challenges posed by attribution perplexing, but the law of attribution is complex.69 In other words, even when the originator of a cyber operation is known, it may be unclear whether his or her actions can be deemed to be those of a state as a matter of law such that the state is in violation of a treaty obligation.

Customary International Law in the Cyber Context

Many obstacles lie in the path of customary norm emergence vis-à-vis cyberspace. The requirement of practice over time hinders this process to an extent, but is not fatal because contemporary customary international law appears to countenance relatively rapid crystallisation. A much greater impediment is the visibility of cyber activities. It is difficult to “see” what goes on in cyberspace. Instead, the effects of cyber operations are often all that is publicly observed; in fact, sometimes even the effects are not apparent to the general public. Therefore, it can be difficult to point to a particular state’s cyber practice to support an argument that a norm has emerged. States, including victim states, may be reticent in revealing their knowledge of a cyber operation, because doing so may disclose capabilities that they deem essential to their security. Undisclosed acts cannot, as a practical matter, amount to state practice contributing to the emergence of customary international law.99

Similarly, states will frequently hesitate to offer opinions.

 

General Principles of Law in the Cyber Context

The third formal source of international legal norms cited in Article 38 of the International Court of Justice’s Statute is general principles of law. A complicating factor with respect to this source is that its nature is the subject of some controversy.108 Generally, the term is said to refer to a number of types of legal principles that are: common across domestic legal systems, such as the use of circumstantial evidence;109 evident from the nature of law itself, for instance res judicata (final judgments of a court are conclusive);110 derive from the nature of international law, such as pacta sunt sevanda (“agreements must be kept”);111 and based on fairness, prominent examples being equity112 and estoppel.113

General principles are most likely to become relevant when disputes between states over cyber matters arise. As an example, in the celebrated Chorzow Factory case, the Permanent Court of International Justice held that the breach of an obligation in international law necessarily gives rise to the obligation to make reparations,114 a principle echoed in the International Law Commission’s Articles of State Responsibility.115 Thus, if a state’s cyber operations violate the sovereignty of another state and cause harm, the former will be obligated to make reparations to the latter. Similarly, courts may decide cases in part based on equitable considerations. Such a decision might be appropriate, for instance, in the case of cyber infrastructure which is shared by states.

However, at times a general principle of law may reflect a substantive obligation. The classic example is the International Court of Justice’s identification of the principle that every State shoulders an “obligation not to allow knowingly its territory to be used for acts contrary to the rights of other States.”116 This Manual Rule 5: “A State shall not knowingly allow the cyber infrastructure located in its territory or under its exclusive governmental control to be used for acts that adversely and unlawfully affect other States.”117

Conclusion

Legal norms are but one facet of the normative environment in which cyber operations exist. To suggest that they alone suffice would be folly. After all, there is a scarcity of cyber-specific treaty law and a near total void of cyber-specific customary law on the subject. As a result, recourse must be had to general international law and the interpretation thereof in the cyber context. Of course, any interpretive endeavour is plagued with uncertainty and ambiguity, especially when engaged in with respect to novel activities such as cyber operations. This lack of legal normative clarity invites states to take differing interpretive positions. A state’s objective view of the law may drive the legal position it adopts; however, it would be naïve to deny that policy and ethical influences have an effect on such determinations.

Controversy and inexactitude will surely characterise this process, which will be neither linear nor logical. The weakening of the early Russian and Chinese objections to the application of extant international law to cyberspace is a milestone in this regard. Yet, while both states have backed away from their opening stance on the issue, it remains unclear where they stand today. Other states such as the United States and the Netherlands are beginning to show a willingness to articulate their positions on how current international law applies in cyberspace. Nonetheless, the public pronouncements to date have been vague, probably intentionally so.

Despite the attention that cyber activities have drawn in the past decade, the conclusion of new treaties or the crystallisation of new customary law norms to govern them is doubtful. Opposition from western states is particularly marked to the former, at least.118 Instead, the application and interpretative evolution of existing international law is the most likely near-term prospect. As to customary law, although it may sometimes develop rapidly, “usually customary law is too slow a means of adapting the law to fast-changing circumstances.”119

Consequently, the work of scholars such as the International Group of Experts who prepared the Tallinn Manual, and those who are engaged in the follow-on “Tallinn 2.0” project, is likely to prove especially influential. This dynamic is appropriate since, as noted in Article 38 of the International Court of Justice’s Statute, the work of scholars is a secondary source of law that informs identification and application of primary sources. But this reality is certainly less than optimal, because states, and only states, enjoy the formal authority to make international law. Unless they wish to surrender their interpretive prerogative to academia, it is incumbent upon them to engage with cyber issues more openly and more aggressively.

In this patchwork and nebulous environment, the role of other normative regimes looms large. Only in exceptional circumstances may their dictates cross the international law border. However, where those boundaries are indistinct, common policy or ethical norms may operate to define the outer boundaries of acceptable conduct in cyber space. Because cyber activities are a relatively new phenomenon, policy and ethical norms may serve to carve out more restrictive boundaries than international laws which are designed to constrain the other activities of states. Over time, these non-legal norms may mature through codification into treaty law or crystallise into customary law, such that they formally define the limits of cyber activities. In the meantime, cyberspace will remain an environment of fervent, and often multi-directional, normative development

Keine Kommentare:

Kommentar veröffentlichen