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

This scenario considers a series of malicious cyber operations originating from one State’s territory and targeted against private entities on the territory of another State. In the course of investigation, the victim State opts to penetrate the networks of its adversary without its consent and discovers that the latter’s military personnel was indeed involved in some of the operations. The scenario analyses the degrees of responsibility of the State of origin, the rules it may have breached, and the ability of the victim State to justify its response under the law of countermeasures.

Keywords
DDoS, data deletion, attribution, non-State actors, sovereignty, due diligence, countermeasures

Facts
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).

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.

The attack has 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 as such has however never admitted any relation to the group.

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 attack 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.

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 (incident 3).

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

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

Similar real-world incidents
Shamoon (2012)

Operation Ababil (2012)

2007 cyberattacks on Estonia

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

Attribution to State B
Whether cyber operations conducted against State A may be attributed to State B depends on the following considerations:

State organs and exercise of governmental authority
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 insufficient for the purposes of attribution of the operation to State B. However, it indicates that State B may have been associated with the operation in question. Nevertheless, as soon as the two officers are identified in incident 5, the situation 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 to 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.

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
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.

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). 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. Similarly, 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
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 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.

There is no obligation to publicly provide evidence upon which State A attributes the cyber operation to State B.

Breach of an international obligation by State B
The following options can be looked at:

Sovereignty of State A
The DDoS attacks (incident 1) probably did not amount to a violation of State A’s sovereignty. There were no physical consequences, 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.

Inherently governmental functions were probably not affected by incident 1 – 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.

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

Prohibited intervention by State B
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 an imprecise category 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.

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
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.

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).

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 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.

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”) 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.

State B therefore violated its obligation of due diligence with respect to those harmful cyber activities in the scenario that 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
(See above for the discussion of sovereignty. This analysis is based on the presumption that sovereignty is a primary rule of international law.)

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.

State A might try to justify its actions by invoking countermeasures.

Countermeasures by State A
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. )

We know that State B’s activities in incident 2 and attributed 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. 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.

However, State A does not have the information about the attribution yet. Is State A engaging in countermeasures, if State B’s internationally wrongful act has not been established?

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.

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.

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.

Bibliography and further reading

 * MN Schmitt (ed), Tallinn Manual 2.0 on the International Law Applicable to Cyber Operations (CUP 2017)
 * Etc.