Scenario 05: State investigates and responds to cyber operations against private actors in its territory

__NUMBEREDHEADINGS__ This scenario considers a series of malicious cyber operations originating from one State’s territory and targeting private entities on the territory of another. In the course of investigation, and after failing to receive cooperation from the suspected offending State, the victim State opts to penetrate the networks of the suspected offending State without consent. The victim State thereafter discovers that the suspected offending State’s military personnel was involved in some of the malicious cyber operations. This scenario analyses the rules of State responsibility, including attribution and the degrees of responsibility of the State of origin, the international obligations that may have been breached, and the ability of the victim State to justify its response under the law of countermeasures.

Keywords
DDoS, computer data, attribution, State organs, non-State actors, evidence, sovereignty, prohibition of intervention, due diligence, countermeasures

Facts
[F1] Important commercial entities and financial institutions in State A fall victim to a massive distributed denial of service (DDoS) type of campaign for several days, resulting in down-time costs that amount to millions of EUR (incident 1).

[F2] Later, it is found that some of the computers were infected by malware and that the data, including operating system data, had been damaged or deleted, rendering the computers inoperative (incident 2). Most of the targeted entities have been able to recover the data from back-ups and restore the functionality of the computers, albeit at a substantial cost. No physical damage was caused in the incidents.

[F3] The incidents have been claimed by a group called “State B Digital Army” that has been active on the Internet for some time and is openly critical of foreign policy of State A. The group bears a name clearly referring to State B, which has in the past been allegedly involved in State-run cyber espionage operations against State A. State B has never admitted any relation to the group.

[F4] The technical aspects of the first incident, or rather its lack of sophistication, seem to suggest that the attackers were just an amateur group. By contrast, the scale of the second incident indicates that a State actor could have been involved in some capacity. Also, most of the group’s activities seem to originate from the territory of State B, although the persons involved and their exact location are unknown to State A.

[F5] Considering that States A and B have not had mutual diplomatic relations for many years, that State B is uncooperative in mutual legal assistance requests, and that there is a risk of destruction of evidence by State B, State A decides to remotely access several computers in the territory of State B without State B’s consent, in the course of a criminal investigation by State A’s cyber police unit in coordination with its intelligence service (incident 3).

[F6] During the cyber operation against State B, State A discovers that a minority of the malicious activities indeed originated from the network of the General Staff of the Armed Forces of State B (incident 4). State A is also able to identify some of the individuals responsible for the attacks and consequently issues public arrest warrants for these individuals, two of whom happen to be military personnel of State B serving in cyber intelligence (incident 5).

[F7] State B denies all allegations, stating that the military personnel were acting on their own, outside their working hours, and it decries the “violation of its sovereignty” by State A.

Examples

 * Cyber attacks against Estonia (2007)
 * Sony Pictures Entertainment attack (2014)

Legal analysis
For a general overview of the structure of analysis in this section, see Note on the structure of articles.

[L1] The legal analysis begins with the attribution of the incidents to States A and B, continues with possible breaches of international obligations by State B (sovereignty, prohibition of intervention, due diligence obligation) and State A (sovereignty), and ends with a possible justification for State A's response (countermeasures).

State organs and exercise of governmental authority
[L2] The fact that the operation against State A was partly launched from the governmental cyber infrastructure of State B, as found out in incident 4, is alone insufficient to attribute the operation to State B. However, it indicates that State B may have been associated with the operation in question. An argument for State B's involvement could be made if the network of its General Staff (as opposed to civilian governmental networks) is especially secure and unlikely to be controlled by a third Party and if the malicious activities were substantial and systematic. Even so, there is insufficient information about a substantial or long-term involvement of State B in the activities.

[L3] Nevertheless, as soon as the two officers are identified in incident 5, the situation clearly changes: even though they might have been acting on their own (as State B claimed in the aftermath of the incidents), their conduct would still be attributable to State B. This is because the mere fact that a State organ was acting ultra vires, i.e., in excess of authority or in contravention of instructions, does not remove attribution from the relevant State. It is not necessary to prove that they were getting orders from their superiors, even though such information might be helpful for the purposes of criminal investigation conducted by State A authorities.

[L4] The information available to State A after incident 5 supports only an attribution of the conduct of the military personnel of State B to that State. By contrast, there does not seem to be sufficient evidence to link State B to any of the activities of the so-called “State B Digital Army” and its supporters. That being said, there are several options under the law of State responsibility on the basis of which such conduct might in fact be attributable to State B:

Non-State actors
[L5] Options 2–4 are manifestly inapplicable to the facts of the scenario given that there is no indication that the so-called State B Digital Army was exercising any form of governmental authority or attempting to form a State of its own. Option 5 (acknowledgment and adoption) also does not apply to the present facts because State B vehemently denied any association with the cyber operations against State A.

[L6] The only remaining mode of attribution is that under option 1, which reflects Article 8 of the ILC’s Draft Articles on State Responsibility (i.e., the existence of instructions, direction, or control). The three standards of "instructions", "direction", or "control" can be satisfied by different conduct, but their common feature is the need to establish a form of subordination between the non-State actor and the potentially responsible State.


 * The arguably most stringent factor of "instructions" would require "State B Digital Army" to be factually subordinate to State B, most commonly as an auxiliary, at the specific moment when the State decides to commit the acts in question. In this regard, the mere fact that State B and the "State B Digital Army" may have shared political goals and aims does not suffice for the purposes of attribution of the latter’s conduct to the former.
 * The second standard would be met if an organ of State B "provided the direction pursuant to which the perpetrators of the wrongful act acted". This would imply a continuing relationship of subordination between State B and "State B Digital Army", going the beyond the mere issuance of instructions without follow-up.
 * Finally, the standard of "control" similarly refers to a continuing process of exercising authority by State B over an activity such as a cyber operation – to meet this standard, State B would have to go beyond merely supporting "State B Digital Army" and would have to go on to determine the execution and course of the specific cyber operation in question.

[L7] The mere fact that some of the activities were conducted from both government and private cyber infrastructure in State B is not sufficient. What would need to be established is some form of subordination between the non-State actor and the potentially responsible State. As the scenario does not offer any indication of such subordination, it can be concluded that the available facts do not support the attributability of the conduct of "State B Digital Army" to State B.

Evidence
[L8] In incidents 1 and 2, State A does not have much information to ascertain who was responsible for the cyberattacks. Its decision to gather more evidence by launching a criminal investigation and accessing State B’s networks without asking for permission (incident 3) might be viewed as reasonable with regard to the circumstances (suspected involvement of State B), if there is no better way to obtain more evidence. This can qualify as a countermeasure of State A against State B (see the sections on extraterritorial jurisdiction and on countermeasures below). However, State A must be aware that it might be committing an internationally wrongful act if State B is not responsible for incidents 1-2, even if the hack-back seemed reasonable in the circumstances.

[L9] There is no legal obligation to publicly provide evidence upon which State A attributes the cyber operation to State B, even though States sometimes do this as a matter of good policy.

Sovereignty of State A
[L10] A violation of sovereignty on the basis of option 1 highlighted above is manifestly inapplicable as there is no indication that either incident 1 or 2 were perpetrated by entities physically present within the territory of State A.

[L11] Inherently governmental functions (options 4 and 5) were probably not affected by either incident 1 or 2 – the inability to pay invoices might have some effect on the functioning of some government organizations in State A, but the causation is probably not proximate enough given the short timeframe of the incident (several days). The organizations can postpone the payments due to unforeseeable and irresistible circumstances, which likely mitigates any adverse effects on their functions.

[L12] As for the remaining options 2 and 3, the DDoS attacks (incident 1) probably did not amount to a violation of State A’s sovereignty. There was no physical damage or injury, even though some individuals and entities in State A could be ‘physically’ affected by being unable to pay for goods and services, and there were significant economic losses as a result of the attacks. The loss of functionality was only temporary and did not require either the replacement of physical components or the reinstallation of the operating system.

[L13] However, the deletion of data (incident 2), leading to a loss of functionality requiring the reinstallation of operating systems, could qualify as a violation of State A’s sovereignty under option 3 above. This is regardless of the fact that the affected infrastructure was non-governmental.

Prohibited intervention by State B
[L14] It would be difficult to argue that State B was involved in a prohibited intervention in State A’s internal affairs. To begin with, it is uncertain whether State A’s domaine réservé — in itself a category of relative scope under international law — was affected by incidents 1-2. The cyber activities were targeting State A’s financial sector, which is traditionally perceived as private, even if State A had a share in its ownership. Perhaps if the operations had been directed at State A’s central bank or at the integrity of financial data upon which State A’s pension or welfare system relied, this would support the qualification of the operation as an intrusion into the domaine réservé.

[L15] Even assuming that the first prong of the test was met by the cyber operations in question, there is no evidence that any of the incidents were coercive in nature as against State A. In other words, the available facts do not support any attempt to compel State A to take any action it would otherwise not take. As such, the cyber operations in the scenario do not seem to qualify as a prohibited intervention under international law.

Due diligence obligation of State B
[L16] The due diligence analysis is important in the situation when State A does not have enough evidence to attribute incidents 1-2 to State B directly. If State B is in breach of its obligation of due diligence, State A might still have the option of justifying its responsive operation (incident 3) as a countermeasure.

[L17] Following from the above analysis of sovereignty, incident 2 (the deletion of data leading to a loss of functionality) is contrary to the rights of State A, and would have been unlawful if conducted by State B. It would be more difficult to similarly qualify incident 1 (the DDoS attack).

[L18] The cyber activities leading to incident 2 were conducted from the cyber infrastructure in the territory of State B; however, State B’s due diligence obligation is not breached solely by the fact that these incidents happened, even though they may have resulted in serious adverse consequences and were contrary to the rights of State A. State A would have to prove that State B had an actual or constructive knowledge of the harmful cyber activities at the time they were launched, and that it neglected its duty to terminate them.

[L19] The information that some of the harmful cyber activities were launched from State B’s government infrastructure is available to State A from incident 4 onwards. Even if it cannot be proved that State B actually gave orders to its organs, or instructed or directed the non-State actors (“State B Digital Army”) to conduct the DDoS attacks and data deletion (see the section on attribution above), the constructive knowledge requirement (“should have known”) likely triggers the breach of its due diligence obligation for the activities originating from its government cyber infrastructure. State B could argue that its government infrastructure was taken over by non-State actors or a third State, or that it did what was to be expected from a reasonable State to terminate the activities, but the burden of proof would then shift to its side.

[L20] State B therefore violated its obligation of due diligence with respect to those harmful cyber activities in the scenario which made use of its governmental cyber infrastructure but could not be directly attributed to it. With respect to the conduct that is attributable to State B, that State bears direct responsibility, which supersedes any violation of due diligence.

Sovereignty of State B: extraterritorial jurisdiction by State A in State B
[L21] As State A decides to remotely access several computers in State B’s territory in search of evidence (incident 3), it is exercising its enforcement jurisdiction in State B’s territory. Absent State B’s consent or other justification, State A’s action is in violation of State B’s sovereignty (option 5 - usurpation of inherently governmental functions).

[L22] State A might try to justify its actions by invoking countermeasures, as detailed below.

Countermeasures by State A
[L23] An interesting question arises at this point: is State A engaging in countermeasures against State B by exercising its enforcement jurisdiction in State B’s territory? (Let us suppose that it would be a violation of State A’s obligation to respect State B’s sovereignty, unless it is State A’s lawful countermeasure. )

[L24] We know that State B’s activities in incident 2 which can be attributed to it based on the information obtained in incidents 4-5 may amount to a violation of State A’s sovereignty, and hence an internationally wrongful act by State B; accordingly, State A would be entitled to respond with countermeasures (supposing that the effects of State B’s activities were still ongoing – for instance, if some of the deleted data could not be restored). Also, let us suppose that State A did not have the obligation to call upon State B for cessation, non-repetition, and reparation, and that it did not have the obligation to notify State B of the impending countermeasure, because that would likely defeat its purpose.

[L25] The countermeasures are not engaging a peremptory norm, nor are they affecting an obligation which would be subject to a dispute settlement procedure. The envisaged countermeasure is likely commensurate with the injury suffered. As for the purpose of the countermeasure, it serves to induce State B to comply with its obligations, because it aims to establish a firm attribution of its internationally wrongful act.

[L26] However, State A does not have the information necessary for the attribution right after incidents 1 and 2. Is State A engaging in countermeasures, if State B’s internationally wrongful act has not been established?

[L27] Ex post facto, it seems that the countermeasures were lawful, because State B had committed an internationally wrongful act; ex ante, however, State A could not have known that. Intuitively, it may seem that State A paradoxically committed an internationally wrongful act in response to an internationally wrongful act of State B.

[L28] However, the standard for the taking of countermeasures is an objective one. The law of State responsibility does not require a subjective element to establish the wrongfulness of an act, nor does it deal with factually impossible attempts of internationally wrongful acts.

[L29] According to the commentary to the Articles on State Responsibility, “[a] State which resorts to countermeasures based on its unilateral assessment of the situation does so at its own risk and may incur responsibility for its own wrongful conduct in the event of an incorrect assessment.” Therefore, State A’s countermeasures were lawful: they were done in reaction to an objectively internationally wrongful act of State B.

Checklist

 * Attribution:
 * How much evidence exists tying the purported perpetrators of the incidents to State B?
 * What should be the evidentiary standard for making the attribution?
 * Does it matter that the operation against State A was partly launched from State B's governmental cyber infrastructure?
 * Does it matter that the operation against State A was partly conducted by public officials of State A?
 * Can the conduct of "State B Digital Army" be attributed to State B?
 * Sovereignty of State A:
 * When does a cyber operation against non-State actors violate the sovereignty of a State?
 * Prohibition of intervention:
 * Did State B's operation intrude into State A's domaine réservé?
 * Was State B's operation coercive?
 * Due diligence:
 * Did the relevant cyber operation adversely affect the rights of State A?
 * Was the cyber operation conducted from or through the territory of State B?
 * Would the cyber operation have been unlawful if conducted by State B?
 * Did the incident result in serious adverse consequences for State A?
 * Did State B know or should it have known of the incident in question?
 * Did State B take all feasible measures to put an end to the malicious cyber activities?
 * Sovereignty of State B:
 * Can State A exercise its jurisdiction in State B's cyber infrastructure when trying to ascertain the attribution to State B?
 * Countermeasures by State A:
 * Does State A commit an internationally wrongful act by responding to an act whose wrongfulness has not been ascertained?

Bibliography and further reading

 * Luke Chircop, ‘A Due Diligence Standard of Attribution in Cyberspace’ (2018) 67 ICLQ 643.
 * Gary P. Corn and Robert Taylor, ‘Sovereignty in the Age of Cyber’ (2017) 111 AJIL Unbound 207.
 * James Crawford, Brownlie's Principles of Public International Law (OUP 2012).
 * James Crawford, State Responsibility: The General Part (CUP 2013).
 * Robin Geiss and Henning Lahmann, ‘Freedom and Security in Cyberspace: Shifting the Focus Away from Military Responses Towards Non-Forcible Countermeasures and Collective Threat-Prevention’ in Katharina Ziolkowski (ed), Peacetime Regime for State Activities in Cyberspace (NATO CCD COE 2013).
 * James Green, ‘Fluctuating Evidentiary Standards for Self-Defence in the International Court of Justice’ (2009) 58 ICLQ 163.
 * Wolff Heintschel von Heinegg, 'Territorial Sovereignty and Neutrality in Cyberspace' (2013) 89 Int’l L. Stud. 123.
 * Joanna Kulesza, Due Diligence in International Law (Brill Nijhoff 2016).
 * Kubo Mačák, ‘Decoding Article 8 of the International Law Commission’s Articles on State Responsibility: Attribution of Cyber Operations by Non-State Actors’ (2016) 21 JCSL.
 * Peter Margulies, 'Sovereignty and Cyber Attacks: Technology's Challenge to the Law of State Responsibility' (2013) 14 MJIL.
 * Tim Maurer and Michael Schmitt, ‘Protecting Financial Data in Cyberspace: Precedent for Further Progress on Cyber Norms?’ Just Security, 14th August 2017.
 * Michael N Schmitt (ed), Tallinn Manual 2.0 on the International Law Applicable to Cyber Operations (CUP 2017).
 * Michael N Schmitt, 'Virtual Disenfranchisement: Cyber Election Meddling in the Grey Zones of International Law' (2018) 19 ChiJIntlL 30.
 * Michael N Schmitt and Liis Vihul, ‘Respect for Sovereignty in Cyberspace’ (2017) 95 Tex L Rev. 163.
 * Sean Watts & Theodore Richard, 'Baseline Territorial Sovereignty and Cyberspace' (2018) 22 Lewis & Clark L. Rev. 771.
 * Rüdiger Wolfrum, ‘Obligation of Result Versus Obligation of Conduct: Some Thoughts About the Implementation of International Obligations’ in Mahnoush H Arsanjani et al, Looking to the Future: Essays on International Law in Honor of Michael Reisman (Brill 2010).
 * Katja Ziegler, “Domaine Réservé”, in Rudiger Wolfrum (ed), Max Planck Encyclopedia of Public International Law (OUP 2008).

Contributions

 * Scenario by: Taťána Jančárková & Tomáš Minárik
 * Analysis by: Tomáš Minárik
 * Reviewed by: Russell Buchan; Dan Efrony; Tomáš Morochovič; Peter Stockburger