Double harm to voters: data-driven micro-targeting and democratic public discourse

Introduction

Online communication – especially on social media – has offered new opportunities for all types of communication. However, among the communicative actions observed, the strategic actions (Habermas, 1984, p. 86) are developing more rapidly than the genuine communicative actions (ibid., pp. 86-101). Political micro-targeting relies on the sophisticated psychological and technological methods, developed by the commercial advertising industry, of collecting information about users' preferences and organising them into user profiles to target them with personalised messages (Papakyriakopoulos, Hegelich, Shahrezaye, & Serrano, 2018; Chester & Montgomery, 2017; Madsen, 2018).

Political micro-targeting can be used for various purposes, directly or indirectly related to political processes: to persuade voters, to (dis)encourage election participation, or donations (Bodó, Helberger, & de Vreese, 2017). It typically involves monitoring people’s online behaviour, aggregating personal data purchased from data brokerage firms, creating extensive databases on voters' personal information, and using the collected and inferred data to display individually targeted political advertisements, sometimes through social media bots (political bots) and other innovative communication methods (Bennett, 2016; Dommett, 2019; Dobber et al., 2019). This paper focuses primarily on political advertisements, which are directly intended for the voters, with political content which may either directly or indirectly inform the voter about a political party's or candidate's opinion, plans or policy; which may invite voters to events and actions, promote causes or incite various emotions.

Political micro-targeting has been with us in the age of old technology, through local campaign meetings, leaflets, door-to-door campaigning (Lupfer & Price, 1972; Devlin, 1973; Kramer, 1970; Keschmann, 2013). But the possibilities of new technology and big data have opened a new dimension (Mayer-Schönberger & Cukier, 2013; Baldwin-Philippi, 2019; Howard, 2006). The 2016 US presidential election, the Brexit campaign and the following national elections in several countries, especially in India and Brazil have stirred great controversies around this campaigning tool, and urged many scholars to examine the political and regulatory implications of the topic (Zuiderveen Borgesius et al., 2018; Howard et al., 2018; Dobber et al., 2017; Evangelista & Bruno, 2019).

Schumpeter describes the democratic process as a competition on the political market (Schumpeter, 2008, pp. 251-268), but there are profound differences between the political and the commercial market.1 Political competition culminates in one common decision passed by the political community, which affects each member of that polity. An open public discourse (Habermas, 1996), and a free exchange of thoughts on the marketplace of ideas (Mill, 1863) would be indispensable in all forms of democracies – whether liberal, competitive, participatory, representative or direct, but most crucial in deliberative democracies. Let me, for the purpose of this article, compare the voters to a jury in a courtroom or in a song contest, who equally participate in the decision. Would it be acceptable in a song contest if the performers sang separately to individually chosen members of the jury, tailoring their performance to address their individual sensitivities? Or, would we accept a suspect and attorney ‘targeting’ the jury in a separate room one by one, based upon their personal characteristics? Political communication is, and should be, more interactive than a show, or a trial, but the main similarity is that the political discourse should be equally accessible to all members of the political community, no niche markets should be developed in the marketplace of political ideas, and no private campaigning performed to specific voters. “In a well-functioning democracy, people do not live in [an] echo chamber or information cocoons” (Sunstein, 2007).

Privacy concerns are in the main streamline of arguments debating political micro-targeting. It is generally accepted that political micro-targeting threatens individual rights to privacy (Bennett, 2013; Bennett, 2016; Kruschinski & Haller, 2017; Bennett, 2019). The rules on data protection require meticulous actions from the advertisers, but it is possible to publish micro-targeted advertising lawfully. It is beyond the purpose of this paper to discuss in more detail all necessary steps to respect personal data protection rules (see e.g., Dobber et al., 2019). This paper focuses on another, yet less considered aspect of related human rights, and therefore refrains from discussing privacy and data protection.

My argument is that political micro-targeting causes double harm to voters: it may violate the rights of those who are targeted, but even more importantly, it may violate the right to information of those who are not targeted and therefore not aware of the political message that their fellow citizens are exposed to. Neither do they have the meta-information that their fellow citizens access, which is the case when, for example, a reader reads a headline but chooses not to read further. In this latter case, the citizen is aware about the information being ‘out there’ and accessible, and has the epistemological knowledge that this piece of information is also part of the public discourse. She has the possibility to read it later, or to ask her friends about the content of the article. She can even listen to discussions among her fellow citizens about the information. But if she is deprived of all these activities as a result of not being targeted with a targeted ad, she suffers harm. "The reason this is so attractive for political people is that they can put walls around it so that only the target audience sees the message. That is really powerful and that is really dangerous." (Howard, 2006, p. 136).

This violation of informational rights could be remedied partly by providing the possibility for the citizens to ‘opt in’, that is, to proactively search and collect the targeted advertisement from online repositories. This remedy is significantly weaker because whether a voter is able and likely to do so, would largely depend on the personal attitudes and characteristics of the voter, leaving especially the vulnerable population in disadvantage.

As those non-targeted may be large parts of society, this violation can be regarded as a mass violation of human rights, a systemic problem which must be addressed by regulatory policy. And in yet another perspective, the practice of political micro-targeting increases the fragmentation of the democratic public discourse, and thereby harms the democratic process. Thus, in my view, this second problem, namely that political micro-targeting deliberately limits the audience of certain content, may prove less curable than the first one. It causes a distortion in the public discourse, which leads to fissures in the democratic process.

Political micro-targeting causes a clash between the two sides of a human right: freedom of expression and the right to access to information. I will argue that the two rights are inseparable, and that political micro-targeting may pose a danger to both of them.

In the first section, I will approach my subject from the perspective of the right to information in the light of the relevant cases of the European Court of Human Rights (ECtHR). In the second section, I will discuss the ECtHR practice on political advertising, and point at an analogy with political micro-targeting. In the third section I argue that even if just a small part of all political micro-targeted advertisements may be manipulative, their impact may be damaging to democracy. Risks shall be assessed as the product of likelihood and impact, and the risk of manipulative microtargeting is one which should not be taken by democracies.

To be able to focus entirely on my argumentation, I will leave the aspect of privacy rights aside - even though it could be considered as the right of the non-targeted voters as well, whose personal data may have been considered, but found unsuitable to be a potential target of micro-targeted political advertisements.

Freedom of expression and freedom of information

There are several justifications for the protection of free speech (the search for truth, the control of power, the constitutive and the instrumental justifications), which are not mutually exclusive (Dworkin, 1999, p. 200; Mill, 1863, pp. 50-58). In the context of this article, the instrumental theory is particularly relevant. According to this theory, freedom of speech is instrumental in inspiring and maintaining a free, democratic public discourse, which is indispensable for voters to exercise their electoral rights in a representative democracy (Baker, 1989; Barendt, 2005, pp. 19-20). Meiklejohn held that the main purpose of free speech is for citizens to receive all information which may affect their choices in the process of collective decision-making and, in particular, in the voting process. "The voters must have it, all of them." (Meiklejohn, 2004, p. 88).

In this regard, individual freedom of expression is a means to reach a social end – the free discussion of public issues, which is a democratic value itself (Sadurski, 2014, p. 20). The public discourse ideally represents a diversity of ideas, and is accessible to a diversity of actors inclusive of all social groups. While mistakes and falsities also form part of the genuine statements expressed by citizens (Meiklejohn, 2004, p. 88), open discussion contributes to their clarification.

On this ground, I would like to show that the right to receive information and the right to freedom of expression are mutually complementary. One cannot exist without the other – this is demonstrated by their listing in the same article both by the European Convention on Human Rights (ECHR, Article 10) and the International Covenant of Civil and Political Rights (ICCPR, Article 19). When the right to receive information is violated, it is freedom of expression in its broader sense, which is violated. While the act of micro-targeting political advertisements realises the free expression rights of the individual politician, at the same time, it harms other citizens' right to receive public information. By depriving non-targeted citizens from the information in the advertisement targeted to others, the act of micro-targeting causes a fragmentation to the public discourse (Zuiderveen Borgesius et al., 2018; see also Howard, 2006, pp. 135-136.), which is an inherent foundation of the democratic process. Therefore, the adverse effect discussed in this article impacts at two levels: at the level of the individual's right to information; and at the collective level of the political community, by disintegrating the public discourse.

The right to freedom of expression is a cornerstone of democracy and a root of many other political rights. Political expression enjoys the highest level of protection; there is little scope for restrictions on political speech or on debate of questions of public interest, as expressed in several judgements of the ECtHR (among others: Lingens v. Austria, 1986, para. 42; Castells v. Spain, 1992, para. 43; Thorgeir Thorgeirson v. Iceland, 1992, para. 63). The margin of appreciation of the member states is narrow in this respect. It is also unquestionably held that the Convention protects not only the content of information but also the means of dissemination, since any restriction imposed on the means necessarily interferes with the right to receive and impart information (where under "means of dissemination", the technological means of transmission were understood, Autronic AG v. Switzerland, 1990; Öztürk v. Turkey, 1999; Ahmet Yildirim v. Turkey, 2012).

Political advertising is also highly protected (VgT v. Switzerland, 2001; Vest v. Norway, 2008), but not entirely limitless (Animal Defenders v. UK, 2013). Section 2 of this paper will discuss the latter decision in more detail.

Apparently, the protection of political expression is rock solid, as it should be in all democracies. And still, against this backdrop, I will argue that the method of political micro-targeting should be regulated, because, in my hypothesis, it violates the right to receive information. In the following paragraphs I will therefore explore the right to receive information in the practice of the ECtHR.

The right to receive information is the passive side of freedom of expression, as expressed both by Article 10 of ECHR, and Article 19 of the ICCPR. The text of Article 10 says: "This right shall include freedom to (...) receive and impart (...) information and ideas", whereas Article 19. ICCPR is somewhat more explicit: "Everyone shall have the right to (...) freedom to seek, receive and impart information and ideas". The ECHR lacks the word "seek", which was observed by the Court (Maxwell, 2017), however, the Court also noted that there was a "high degree of consensus" under international law that access to information is part of the right to freedom of expression, as shown by the relevant decisions of the UN Human Rights Committee regarding Article 19 of the ICCPR. 

The ECtHR practice shows a tendency of growing recognition of the right to receive information, as shown by Kenedi v. Hungary (2009), Társaság a Szabadságjogokért v. Hungary (2009), Helsinki v. Hungary (2016). This is a clear development from the Court's previous attitude where it denied that the right of access to information fell within the scope of Article 10 (Leander v. Sweden, 1987; Gaskin v. UK, 1989; Guerra v. Italy, 1998). For example, in Leander v. Sweden, the Court held that the right to freedom of expression in Article 10 did not confer a positive right to request information (Maxwell, 2017). But in a report issued by the Council of Europe - already before the cases of Kenedi v. Hungary and Társaság v. Hungary, the author emphasised that "The ambit of freedom of information thus has a tendency to expand: this freedom is particularly important in political or philosophical discussion, given its role in helping to determine people’s choices" (Renucci, 2005, p. 25).

In Társaság v. Hungary, the Court declared that itself “has recently advanced towards a broader interpretation of the notion of ‘freedom to receive information’ (see Matky v. la République tchèque, 2006) and thereby towards the recognition of a right of access to information.” In this case, the Hungarian non-governmental organisation (NGO) "Civil Liberties Union" asked for access to a petition submitted by a member of the parliament to the Constitutional Court, which questioned the constitutionality of newly passed amendments to the Criminal Code, related to drug abuse. The NGO which is active in the protection of human rights, and which had been working in the field of harm reduction of drug abuse, was specifically interested in that topic. The Constitutional Court denied access to the petition without the approval of the petitioner. The national courts both approved this decision, referring to the protection of personal data of the petitioner. ECtHR noted that this case was related to interference with the watchdog function – similar to that of the press, rather than the violation of the general right to access to public information. It added that the obligation of the State includes the elimination of obstacles which would hinder the press to exercise its function, if these obstacles exist solely because of the information monopoly of the authorities – as in this case the information was ready and available. Therefore, the Court established violation of Article 10 of ECHR, because such obstacles to prevent access to public information can discourage the media or similar actors to discuss such issues, and consequently they would be unable to fulfil their "public watchdog" role.

In Kenedi v. Hungary (2009), a historian claimed access to documents of the national security services of the communist regime, which were restricted by law. The Court emphasised that "access to original documentary sources for legitimate historical research was an essential element of the exercise of the applicant’s right to freedom of expression". In fact, the subject matter in the conflict reached beyond the historical information, because the restricted documents of the former communist secret service related to persons still actively working, and had the potential to stir substantial political controversy. Access to the information thus could contribute to a free political debate. The Hungarian courts judged in favour of the researcher Kenedi, but their decision was not executed by the government. This failure made the case so clearcut, that the Court did not go into particular detail in the argumentation section.

The mentioned cases were instances where the state's reluctance to reveal public information impaired the exercise of the functions of a public watchdog, like the press, or an NGO, which intended to contribute to a debate on a matter of public interest (Helsinki v. Hungary, 2016, para. 197, and Társaság v. Hungary, 2009, para. 28). The Court previously had expressed that preliminary obstacles created by the authorities in the way of press functions call for the most careful scrutiny (Chauvy and Others v. France, 2004 - cited in Társaság v. Hungary, 2009, para. 36.). The Court also considered that obstacles created in order to hinder access to information of public interest may discourage those working in the media or related fields from pursuing such matters (citing Goodwin v. the United Kingdom, 1996, para. 39 in Társaság v. Hungary, 2016, para. 38).

This argumentation could be mutatis mutandis relevant if access to political advertisements during or after an election campaign would be restricted only to targeted voters, as the scrutiny exercised by NGOs and journalists as well as election authorities over the election campaign is an inherent part of their watchdog role, to ensure and supervise the fairness of elections. In the mentioned cases, the obstacle in the way of access to information were created or maintained by governments or state bodies (such as the Constitutional Court, or police departments). However, in many other instances, the Court decided in favour of freedom to receive and impart information against the interests of private enterprises (Bladet Tromsø v. Norway, 1999; Sunday Times v. UK, 1979). In these and other cases, the Court emphasised that the right to access to information is not reserved to the press: the general public is also entitled to access public information (De Haes & Gijsels v. Belgium, 1997, para. 39; Fressoz & Roire v. France, 1999, para. 51).

In all the cited cases, the Court had to decide between a restriction of Article 10 for some legitimate interest. Freedom of information and freedom of expression were mutually completing each other, freedom of information being instrumental to freedom of expression. The applicants' right to receive information was violated which prevented them in exercising their right to freedom of expression.

Political micro-targeting represents a specific niche category among the political advertisements –– although there is some debate about what the term actually includes, and is relatively new to the European jurisprudence. Therefore, to this date, there has been no case at the ECtHR related to political micro-targeting. Nevertheless, it should be noted that the general public interest has always been an important factor in finding the balance between colliding rights, as an official Council of Europe report states: "In determining whether or not a positive obligation exists, regard must be had to the fair balance that has to be struck between the general interest of the community and the interests of the individual, the search for which is inherent throughout the Convention.” (CoE, 2005, p. 42). In addition, from the entirety of the case law of the ECtHR, it can also be deducted that in balancing the restriction of freedom expression the free public debate of matters of public interest has been considered with decisive weight (Sunday Times v. UK, 1979; Bladet Tromsø v. Norway, 1999).

Political advertising

The scenario in the case of political micro-targeting is somewhat different from the above cases, and more similar to the cases related to political advertising such as Animal Defenders v. UK (2013), and Erdogan Gökce v. Turkey (2014), or TV Vest v. Norway (2008). In Animal Defenders v. UK, an NGO was prohibited from running their public issue ad campaign on television, due to a legal prohibition of broadcasting political advertisements. In TV Vest v. Norway, the broadcasting company was fined for having broadcast the political advertisements of a small and powerless pensioners’ party despite the legal prohibition. In Erdogan Gökce v. Turkey, the applicant, who distributed political campaign leaflets a year ahead of elections, was sentenced to three months of imprisonment.

In these cases, freedom of expression was limited by state intervention with the aim to protect democratic discourse, to ensure equal chances to all political candidates. Thus, access to specific political information was limited by the respective states in order to ensure the right of the general public to receive information in a fair and undistorted way.

Despite the similar factual background, the details and so the outcomes of the cases were different. In Erdogan v. Gökce, the prescribing law was less than clear, and its application had been inconsequential previously. These circumstances of the case set a clear case for a violation of Article 10 of ECHR.

Nevertheless, the Court in all cases assessed whether the applicant's right to communicate information and ideas of general interest - which the public has the right to receive, could be justified with the authorities' concern to safeguard the democratic debate and process, and to prevent it from being distorted during the electoral campaign by acts likely to hinder fair competition between candidates (Erdogan Gökce v. Turkey, 2014, para. 40, citing Animal Defenders, para. 112). In my view, this rationale offers a sound interpretation even of the Animal Defenders judgment, in which the Court found no violation of Article 10 of ECHR which was greeted with perplexity by many commentators (Ó Fathaigh, 2014; Lewis, 2014; Rowbottom, 2013b).

In all of these cases, paradoxically from the perspective of Article 10 of ECHR, freedom of speech was to be restricted with the objective to preserve a sound informational environment; because pluralism of views, and ultimately the democratic process would otherwise have been distorted by the speech in question.

I will below analyse in more detail the Court's reasonings related to political advertising, through the examples of these three landmark decisions (Animal Defenders v. UK, 2013; TV Vest v. Norway, 2008; and Erdogan Gökce v. Turkey, 2014). First, I would like to show that the considerations in TV Vest, which preceded Animal Defenders, have signalled the Court's position which was followed also in Animal Defenders, and therefore, in my view, the latter should not have been as surprising as it was widely regarded. In TV Vest, the Court has carefully considered the government's argumentation that the rationale for the general prohibition of broadcast political advertisements was that such type of expression was "likely to reduce the quality of political debate generally", so that "complex issues might easily be distorted and groups that were financially powerful would have greater opportunities for marketing their opinions than those that were not". Therein, "pluralism and quality were central considerations". The Court accepted this as a legitimate aim of the regulation, but held that the restriction did not qualify the expectations of proportionality, primarily because the applicant Pensioners Party was not a financially strong party, which would have been the targets of the prohibition, on the contrary: it "belonged to a category for whose protection the ban was, in principle, intended" (at 73).

In my interpretation, here the Court suggested that the law's effect is lifted for the sake of a party which was meant to be a beneficiary, rather than one to bear the burden of the prohibition. It was precisely this case-by-case distinction which was distinguished in Animal Defenders, where the Court declared that "the more convincing the general justifications for the general measure are, the less importance the Court will attach to its impact in the particular case". Moreover, the Court explained that "a prohibition requiring a case-by-case distinction(...) might not be a feasible means of achieving the legitimate aim" (para. 122). Thus, here the Court not only accepted the legitimate aim of the prohibition, but also accepted that no exception should be made even for an otherwise socially benign NGO campaign, because the case-by-case application "could lead to uncertainty, litigation, expense and delay as well as to allegations of discrimination and arbitrariness, these being reasons which can justify a general measure" (para. 122).

After examining the similarities of argumentation in the consecutive decisions, I would like to describe how the Court identified the decisive factors in the case of Animal Defenders.

In Animal Defenders, the Court held that the ban's rationale served the public interest: "the danger of unequal access based on wealth was considered to go to the heart of the democratic process" (para. 117); the restriction had strict limits as it was confined to certain media only, and a range of alternative media were available. The Court observed that it needed to balance between the NGO's right to impart information and ideas, which the public was entitled to receive, with the interest of the democratic process from distortion, by powerful financial groups which could obtain competitive advantages in the area of paid advertising and thereby curtail a free and pluralist debate (para. 112). At the same time, the Court acknowledged that both parties had the same objective: the maintenance of a free and pluralist debate on matters of public interest (see also Rowbottom, 2013a).

From this reasoning, we can conclude that political advertisements can be restricted with certain conditions:

• their dissemination would impose a risk of unequal access based on wealth;
• the legitimate aim is protection of the democratic process from distortion;
• the lurking distortion would cause competitive advantages and thereby curtail a free and pluralist debate;
• the restriction has strict limits by being confined to certain media only, and other media is available.

The dictum of Animal Defenders is, that under these conditions, the right of social organisations to impart information and ideas – which the public is otherwise entitled to receive – may be restricted.

Translating this into micro-targeted political advertising, we can recognise similarities: the means to apply this technology is not equally accessible to all political parties (or issue-groups) without regard to financial resources, and the concluding distortion of the public discourse might harm the democratic process, and curtail free and pluralist debate. Thus, we can conclude that, with narrowly curtailed rules, such political advertising can be restricted without violating Article 10 of ECHR.

Now, after having drawn an analogy between the decisions and political micro-targeting, two more questions are to be addressed.

First, both in Animal Defenders, TV Vest and in Erdogan Gökce v. Turkey, the applicant's right to freedom of expression was restricted by the state, and this restriction led the applicant to apply to the Court. Today, micro-targeted political advertisements are not restricted by state regulation. My conclusion implies that in the case this would happen, such restriction would not be against Article 10 of ECHR. As demonstrated in the analysis above, I interpret the Court's judgements that such a restriction would be regarded by the Court as serving a legitimate aim, as necessary in a democratic society, primarily to prevent the public political debate from distortion, and secondarily, on the basis of right-to-information case law (see above), to ensure public scrutiny by non-targeted users, including journalists and NGOs who may not otherwise have access to all the political advertisements. Its proportionality is naturally dependent on the nature of the specific regulation, but one of the main aspects would be confinement to a certain media type only, so that other means of publicity remain available.

Second, the fact that a state restriction would not be contrary to Article 10 of ECHR, does not mean that such a restriction is necessary to preserve the sound, democratic public discourse. The Court has declared in its previous decisions (Leander v. Sweden, 1987; Gaskin v. UK, 1989, para. 57; Guerra and Others v. Italy, 1998, para. 53; and Roche v. UK, 2005, para. 172) that Article 10 of ECHR does not grant an entitlement to demand that the state actively ensures access to all public information.

Thus, by saying that state restriction of political micro-targeting would be acceptable from the perspective of human rights, I did not yet prove that it is also desirable. In the following section I will discuss the effect of micro-targeting on the democratic process, and its further consequences.

The risks of political micro-targeting to democracy

Evidence shows that political micro-targeting can increase polarisation and fragmentation of the public sphere. For example, in the US presidential campaign of 2016, Facebook posted ‘dark posts’ (sponsored Facebook posts that can only be seen by users with very specific profiles) to micro-target groups of voters with “40-50,000 variants of ads every day” (Illing, 2017), among other reasons, to discourage voters (see also Chester & Montgomery, 2017). In a negative scenario, when political parties share only those fragments of their political programmes with the targeted voters who such programmes would likely support, and other fragments with yet another part of the audience, that can be harmful for democratic processes (Zuiderveen Borgesius et al., 2018; see also Howard, 2006, pp. 135-136.). Beyond being an unfair practice, this splinters the shared information basis of society and contributes to a fractured public sphere.

However, technology in itself is neither evil, nor good. The strong potential of micro-targeting could also serve the public interest, if applied purposefully for that end. There is ample research and discussion on how social media engagement enhances democracy (Sunstein, 2007; Sunstein, 2018; Kumar & Kodila-Tedika, 2019; Martens et al., 2018). For example, it could be exceptionally effective in transmitting useful messages to citizens on healthy living, safe driving, and other social values with which it can greatly benefit society. In this perspective, data-driven political micro-targeting has the potential to increase the level of political literacy and the functioning of deliberative democracy, by incentivising deliberative discussion among those voters who are interested and who feel involved. However, even in this case, the non-targeted citizens are excluded from the discussion, without having been offered the choice to participate in it, unless effective measures are used to enable their involvement. At the core of the issue is the paternalistic distinction between citizens, deciding on their behalf whether they should get certain information or not (see also Thaler & Sunstein, 2008).

I would like to emphasise that the process of democracy needs to be protected with utmost care. Why? We are witnessing a success of populist political campaigns globally, and we should accept the fact that social representation of ideologies is diverse, and may change over time. It is even possible that popular support for the idea of deliberative democracy will decrease, as democratic processes have signalled in several countries, even within the European Union. But at the same time, there is (still) a global consensus on the universal protection of fundamental rights, democratic processes and the rule of law, which form the fundamental legal structures of our societies. Therefore, while political communication should not be restricted on the basis of its content, even ideologies which are critical of the current forms of democracy (e.g., illiberalism) should be allowed to compete in the political arena, but it must be secured that all democratic discussion and political battles in which these ideologies wrestle are played under fair circumstances. Only this can ensure that the freedom of political expression is not abused and the democratic process is not hacked by political opportunism. Political campaigning is one of the key processes by which the formation of the democratic will of the people is generated, and should this process violate fundamental rights, that would in itself pose a threat to democracy. It would destabilise the democratic process and raise issues of legitimacy, whether or not the controversial technique is successful. Respect for fundamental rights is a prerequisite for the rule of law, and the rule of law is a precondition for democracy. The three are “inherently and indivisibly interconnected, and interdependent on each of the others, and they cannot be separated without inflicting profound damage to the whole and changing its essential shape and configuration” (Carrera, Elspeth, & Hernanz, 2013).

In sum, I argue that online micro-targeting should be restricted not because it can carry manipulative content, and not because it can violate privacy rights, but because it threatens the process of democratic discourse. Even if the likelihood of manipulation is small, the harm that can be caused is so severe that the overall sum of the risk is too high to be taken. Direct political campaigning has been with us before, in the form of door-to-door canvassing, leaflets, local meetings, and other tools. However, access to masses of voters’ personal data, the analysis of these databases with advanced technology, and the low cost of personalised communication generate a qualitatively new situation. The voters have lost their control over being targeted, and the transparency of the targeting has diminished (see also Mayer-Schönberger & Cukier, 2013; Baldwin-Philippi, 2019; Howard, 2006).

Having argued above that the technique of micro-targeting is harmful to the individual right to information, and that it threatens the collective democratic public discourse, the logical conclusion would be to recommend a complete prohibition of using this strategic communication tool in the political discourse. Only this could eliminate the risk to the informational rights of masses of voters and to the further polarisation of the public discourse.

However, considering also the benefits and the high interest of the political elite in this tool, the political reality is likely to incline towards allowing its use and demanding appropriate safeguards – the discussion of which is beyond the limits of this article.

Conclusion

This article argued that micro-targeting violates the fundamental right to receive information, and the collective right to the public discourse. Thereby it harms the democratic process of deliberation. Non-targeted voters' right to receive information is violated by being excluded from political communication that is supposed to be public and inclusive in a democracy. This is a mass violation of a human right, which is part of the right to freedom of expression, as recognised by the ECtHR. To focus entirely on the informational rights of non-targeted citizens, the article avoided the discussion of other rights that may be affected.

I examined two aspects of the ECtHR jurisprudence in freedom of expression: the right to receive information, and freedom of political expression. In the first topic, I showed that in many instances, the Court decided in favour of freedom to receive and impart information for the sake of the public discourse, even against the interests of private enterprises (Bladet Tromsø v. Norway, 1999; Sunday Times v. UK, 1979). I demonstrated that the right to receive information is not reserved to the press, but it includes the general public as well (De Haes & Gijsels v. Belgium, 1997, para. 39; Fressoz & Roire v. France, 1999, para. 51). In our context, the right to receive all political information is regarded as crucial for non-targeted voters, including journalists, NGOs and election authorities. Although the above cases relate to restrictions caused by the state rather than private entities, the Court found that the state is obliged to eliminate obstacles which would hinder the press to exercise its watchdog function (Társaság v. Hungary, 2009). Whenever the Court had to balance between the public interest of the community and the interest of an individual, the public interest has been considered with substantial weight (CoE, 2005, p. 42; Sunday Times v. UK, 1979, Bladet Tromsø v. Norway, 1999).

In my analysis of the ECtHR decisions relating to political advertising, I show that the Court had consistently found acceptable restrictions of political advertising in the interest of the sound democratic public discourse. I argue that even though the Animal Defenders v. UK (2013) decision was regarded as exceptional then, but both preceding and following judgments clearly show the consistency of the Court’s position. In all discussed cases, the Court assessed the right to political expression and to receive information versus the protection of the public discourse, where the latter was considered as the authorities' responsibility to prevent the democratic debate from being distorted (Erdogan Gökce v. Turkey, 2014; Animal Defenders v. UK, 2013, para. 112). In the prior TV Vest v. Norway case (2008), it was shown that the Court accepted the principle that a certain type of political speech threatened with reducing the quality of the political debate, and causing distortion of the discussion, as well as inequality between the financially powerful and less well-financed groups, even though in the specific case the Court found the restriction unproportionate, because the party in question was a small and financially weak party (TV Vest v. Norway, 2008). In the case of Animal Defenders, the Court found acceptable the restriction of the dissemination of public issue ads with the objective to preserve a sound informational environment. The factors which the Court identified so as to determine the proportionality of a restriction can be guiding for the case of political micro-targeting as well: to prevent the risk of unequal access to the public discourse based on wealth, and consequently the protection of the democratic process from distortion, which would curtail a free and puralist debate. In Animal Defenders the restriction was found to be sufficiently narrowly tailored, as it applied to certain media only, and other media remained available. These arguments also apply to the case of online political micro-targeting, which is suspected to fragment and distort the democratic process through the informational deprivation caused to non-targeted voters.

While discussing policy options is beyond the constraints of the article, to conclude, here are a few thoughts to consider.

First, some countries' legal culture may incline towards more risk-taking, even at the price of certain collective and individual rights being harmed, whereas others are more risk-averse. Similar to the dispute over hate speech, the former culture would rather tolerate the risk to the political process, than restrict individual freedom of expression. Long-standing, stable, and prosperous democracies may find the explained risks more manageable – this would be more characteristic to the United States than to most member states of the European Union (Heinze, 2005; Kahn, 2013).

Second, staying with the example of hate speech, while there are important differences between the European member states in regulating hate speech, the similarities are more characteristic, especially in contrast to the United States. Importantly, the ECtHR held that the margin of appreciation is narrow in the field of political freedom of expression. Legislative efforts also have to face the difficulties of the transborder nature of targeting and advertising (Bodó, Helberger, & de Vreese, 2017). All these factors highlight the importance of an international, but at least EU-wide policy approach (see also Dobber et al., 2019).

Third, when it comes to the protection of fundamental rights, states have an obligation to ensure that these rights are not restricted even by private entities. Self- and co-regulation does not impose sanctions in case of non-compliance, therefore they do not provide sufficient protection for individual human rights. For political advertisers, the stakes are higher than in any other industry, and these circumstances render the long-term success of self-regulation less likely. Specifically, in the case of political micro-targeting, the data controllers are political parties that had, have or are going to have governmental power, and thus have a potential influence on national regulations and on authorities as well. This further raises the significance of supranational regulation and supervision by EU bodies.