Cyber and Influence Operations Targeting Elections: Back to the Principle of Non-Intervention

Written by

After the 2016 US presidential election, President Obama criticized Russia for interfering in the vote, but stopped short of alleging a violation of international law. The intervening period has seen a vigorous debate on the rules governing interference in elections, but no consensus has emerged. The 2020 presidential election would seem, then, as Michael Schmitt reminds us, in a series of posts for this Blog (here, here and here), a good time to revisit the issue, especially in light of reports suggesting that foreign powers are again trying to interfere in the vote, by hacking information and communications technologies (ICTs), disseminating fake news stories, and conducting disinformation campaigns.

The Problem of “Coercion”

The standard way that international lawyers approach the subject of foreign interference in domestic political affairs is through the lens of the non-intervention rule. In its 1986 Nicaragua judgment, the International Court of Justice explained that intervention is wrongful “when it uses methods of coercion” (para. 205). This is believed to create difficulties for the application of the rule to the cyber domain, because coercion is said to require a conscious unwilling act on the part of the victim (as in the case of “Your money or your life”, where the victim consciously hands over the money, albeit they do so unwillingly). But this understanding does not easily translate to cyber, where the main threats are the clandestine hacking of the ICTs used in elections–when the target state is not aware of the hack, and influence operations targeting citizens through social media, so they willingly vote for a different candidate.

Lack of agreement as to whether the hacking of the computers can be categorized as “coercive” has led some to argue we should abandon the non-intervention rule and look instead to the principle of sovereignty to find a rule prohibiting election interference (for a good introduction to the debate, see the Chatham House Research Paper). This is the approach of the influential Tallinn Manual 2.0 (CUP, 2017), which explains that the sovereignty principle differs from the non-intervention rule, because “intervention requires an element of coercion” (Rule 4, Explanatory para. 22). The downsides with the ‘sovereignty’ argument are threefold: first, that the scope of the principle is the subject of significant disagreement; second, it is not clear, as Mike Schmitt concedes, ‘whether all interference with an election is encompassed in the rule’; and, finally, the rule has nothing to say about influence operations targeting voters. Moreover, as non-intervention provides the basis on which states and international lawyers already talk about the limits on intermeddling in domestic political affairs, we should not lightly abandon the principle when looking to explain the rules governing foreign interference in democratic votes.

The Meaning of “Coercion”

To make sense of the cyber non-intervention principle, we need to be clear about the meaning of “coercion.” The government of the Netherlands explains the issue this way:

“The precise definition of coercion, and thus of unauthorised intervention, has not yet fully crystallised in international law. In essence it means compelling a state to take a course of action (whether an act or an omission) that it would not otherwise voluntarily pursue. The goal of the intervention must be to effect change in the behaviour of the target state.”

The quote nicely captures two things: Agreement that coercion describes a situation where the outside power gets the target state to do something it would not otherwise do voluntarily; and the lack of consensus on what “methods of coercion” are covered, especially in the cyber domain. 

Our objective, then, must be to explain which cyber and influence operations can be categorized as “coercive”. To do this, we can look to the long-established philosophical debates on the meaning of “coercion” in inter-personal relations, on the understanding that we use the same term when describing getting a person and a state to do something they would not otherwise do. We see this in the Vienna Convention on the Law of Treaties, which establishes that a treaty is void, if it has been procured by the “coercion of a representative of a state” (Article 51), or by the “coercion of a state” (Article 52).

Hacking elections

Whilst we often think in terms of coercive threats, typically in the form “Your money, or your life”, many philosophers agree the term can also be applied to circumstances of “physical coercion”. In other words, they conclude there is no difference between the scenario when someone says, “Sign this treaty, or I will kill you”, and when they grab your hand and force your signature onto the document. In both cases, you are forced to do something you would not otherwise do.

In its 1986 Nicaragua judgment, the International Court of Justice confirmed that “the element of coercion… is particularly obvious in the case of an intervention which uses… force” (para. 205). There is not a violation of the non-intervention rule simply because the outside power uses force against the target state (where this is the case, the ICJ uses the language of “use of force” and “violations of sovereignty” (para. 251)). The non-intervention rule is concerned with the situation where the target state is made to do something by the outside power. In the Nicaragua case, the complaint was that the USA was trying to “coerce the government of Nicaragua into the acceptance of United States policies and political demands” (para. 239).

Cyber power creates new opportunities for states to force other countries to do something they would not otherwise do, by taking control of their ICTs. In the case of election interference, this includes removing people who have traditionally supported one party from the electoral roll; interferences with electronic voting machines to change voters’ preferences or make votes disappear; or even changing the outcome of the election by hacking the vote tabulation software.  

By taking control of a government’s computers or computer network, the outside power forces the target state to do something that it would not otherwise do, leaving it with no choice in the matter. This kind of cyber power is coercive in the same way that grabbing the hand of a state’s representative and forcing them to sign a treaty is coercive. All cyber operations that take control of the ICTs that underpin the holding of elections are “coercive”, and therefore wrongful under the non-intervention principle, because the outside power achieves its objective by working through government bodies, like the Electoral Commission, compelling them to do something they would not otherwise do.

The Clear Line Between Influence and Coercion

One feature of the Internet is that it allows foreign powers to try and directly influence political discussions in other states, especially by way of social media. Under the non-intervention principle, influence operations are wrongful when they use “methods of coercion in regard to such choices, which must remain free ones” (Nicaragua case, para. 205). There is no doubt that an election concerns a choice that should remain free. The only question is whether cyber influence operations can be categorized as “coercive”.

Again, we can look to the debates on coercion in inter-personal relations to make sense of the “coercive” nature of some inter-state propaganda.

The legal philosopher, Joel Feinberg has shown how the various techniques for getting a person to do something they would not otherwise do can be placed along a “spectrum of force” (Joel Feinberg, Harm to Self (OUP, 1986) 189). So, for example, if P wants Q to stay in a room, P can tell Q that, if they leave the room, P will kill them (coercion); tell Q there is a terrorist outside, with a suicide vest, when this is not true (coercive manipulation); tell Q that P will be upset if Q leaves the room (manipulation); or simply ask Q to stay in the room, because P has a parcel being delivered that day (request for action). In all cases, P’s objective is the same – P wants Q to stay in the room. The division is between P’s actions that leave Q with a meaningful choice, and those which do not, i.e. between influence and coercion.

There is no problem with one person trying to get another person to do something they would not otherwise do – providing this does not undermine their right to decide things for themselves. In our example, P tries to persuade Q to stay in the room by saying “I have a parcel coming today”. P has introduced a new piece of information to try and influence Q’s decision-making, but Q will still make the decision whether to stay in the room, or not.

The position is different where P lies about the facts. By lying, P deceives Q by changing their understanding of the world and the basis on which they make a decision. This is wrong because Q should be able to make up their own mind, based on a proper understanding of the facts. In an important contribution to the philosophical debate on the nature of lying, Hugo Grotius explains that lying is wrong, because it undermines the right of the target to “liberty of judgment” (Hugo Grotius, The Rights of War and Peace (Walter Dunne, [1625] 1901), Book III, Ch I, § XI).

All lies are deceptive, in the sense of deceiving the target about the reality of the situation. But if P chooses the right lie, then P can get Q to act and leave Q without a meaningful choice about what to do. In our example, P said to Q, “Do not go outside, there is a terrorist, with a suicide vest”. If Q believes this lie, they will have been deceived about the reality of the situation and will be certain to stay in the room. P will have made Q do something that Q would not otherwise have done, leaving them without a meaningful choice in the matter. Mike Schmitt makes a similar point when he notes that citizens could be prevented from voting by false social media reports of an active shooter situation, concluding: “Reasonable individuals would follow those instructions, and thus not cast their vote.”

In other words, there are circumstances when lying is the functional equivalent of coercion.

Unwelcome and Unlawful Intermeddling

There is widespread agreement in the literature that “just providing the facts” to the citizens of another country, including information critical of the government, does not constitute a prohibited intervention. In the same way that attempting to influence another person by giving them new information is not wrongful, efforts by one state to influence the population of another by providing new facts and commenting on news stories is not prohibited under the non-intervention doctrine.

The principle that “just providing the facts” is not a violation of the non-intervention rule applies equally to information gained by hacking computer systems and making that information public, with the objective of influencing political debates. Thus, no matter how unwelcome the Russian interference in the UK’s 2019 General Election, by leaking documents on US-UK trade discussions, this cannot be categorized as a violation of the non-intervention principle, because the objective of the alleged influence operation was to place factual information into the public domain.

Fake news does not, by definition, enjoy the protection accorded to factual information under the principle of non-intervention, but there is no specific prohibition on fake news. Fake news is only wrongful, then, where it can be categorized as “coercive.”

Coercion, as we have seen, describes a situation in which the outside power forces the target to do something they would not otherwise do. One way this can be done is by lying to the electorate with the intention of deceiving them into thinking and then voting differently.

Take the example of a foreign power releasing a “deep fake” video, which appears to show, in convincing detail, a candidate for office doing something they never did (engaging in illegal or immoral acts, for example); saying something they never said; or that presents a false picture of reality (deep fakes have been released that question Joe Biden’s health, although there is no evidence this is the work of an outside power).

If citizens vote for another candidate because of a “deep fake”, or some other form of “fake news”, they will have been deceived into doing something they would not otherwise have done. Moreover, they will have been given no meaningful choice in the matter, because they now have a false perception of the true facts of the situation, which formed the basis on which they cast their votes.

Fake news is “coercive” when the communication is intended to deceive the target population into doing something they would not otherwise have done, absent the false information. This is coercive in the same way that lying is coercive when it leaves the target without a meaningful choice about what to do. In other words, where an outside power is responsible for a “fake news” story, including a “deep fake” video, intended to influence the electorate’s decision-making to such an extent that they are left without a meaningful choice about how to vote, that communication is a violation of the principle of non-intervention.

Conclusion

The correct meaning of the term “coercion” in the cyber non-intervention principle will emerge from the things that states say about the practice of foreign state cyber and influence operations. It is, therefore, important for democratic countries to explain why certain operations can be categorized as coercive. This blog has developed an argument for how we can, and should, understand the notion of coercion. The advantage of this approach is that the cyber non-intervention principle outlined here can protect the integrity of the democratic process from election hacking and the most egregious examples of disinformation, without the need to invent new rules and then persuade others of their content and binding nature. The non-intervention rule is accepted by all states, regardless of their political system or foreign policy goals. It therefore provides the basis on which we can agree the binding international law rules on cyber and influence operations targeting elections.

This Blog draws on a forthcoming publication on the meaning of “coercion” in the non-intervention principle in the Duke Journal of Comparative & International Law.

Print Friendly, PDF & Email

Leave a Comment

Your comment will be revised by the site if needed.

Comments

Mike Schmitt says

October 26, 2020

Very interesting and insightful Steven; I look forward to the Duke article. I usually don't respond to pieces, but allow me to clear up one point in yours. You say "Lack of agreement as to whether the hacking of the computers can be categorized as 'coercive' has led some to argue we should abandon the non-intervention rule and look instead to the principle of sovereignty to find a rule prohibiting election interference... This is the approach of the influential Tallinn Manual 2.0, which explains that the sovereignty principle differs from the non-intervention rule, because 'intervention requires an element of coercion'". I'm not sure if you meant we (the Tallinn team) wanted to abandon intervention for that reason OR that we were offering an alternative (interference with an inherently governmental function). If the former, absolutely not. Indeed, we used election meddling as our example. If the latter, absolutely. The fact that election meddling is not coercive doesn't mean it is not a violation of sovereignty, even though much election meddling will violate both primary rules. Just wanted the readers to know where we came out on the issue. Thanks again for your great piece which I enjoyed. Best, Mike