‘No man and no mind was ever emancipated
merely by being left alone.’
4.1 On Dangers and Solutions
On 11 July 2017, the Knight First Amendment Institute at Colombia University filed a lawsuit against former President Trump and his aides for blocking several people from Trump’s Twitter account following their criticism of his presidency and policies. The plaintiff asserted that the @realDonaldTrump account was a ‘public forum’ protected under the First Amendment, from which no one could be excluded based simply on their views.Footnote 1 The US Court of Appeals for the Second Circuit affirmed the district court’s holding that such blocking violated the First Amendment.Footnote 2 This judicial finding confirms ‘the key role that the Internet can play in mobilising the population to call for justice, equality, accountability […] and better respect for human rights’.Footnote 3
By way of contrast, assume now that Twitter has removed users’ comments. The outcome would be different since Twitter, as a private actor, is not bound by constitutional obligations embodied in the First Amendment. Under such a scenario, First Amendment rights are illusory, as are sometimes privacy-related rights, whose online infringements can produce cascade results. Recently, a surveillance information technology company and social media made, what Danielle Citron calls ‘intimate privacy’,Footnote 4 accessible to everyone possessing a computer or mobile phone. MIT Technology Review revealed how iRobot collected photos and videos from the homes of test users and employees and shared them with data annotation companies.Footnote 5 The investigation revealed that images of a minor and a tester on a toilet ended up on Facebook. It was a robot vacuum that took the pictures, not a person. Nevertheless, it would be too simple to blame iRobot for this infringement since humans were those who decided to steal and leak them.
But who is to blame? When during the 1990s the internet started the digital revolution, it was generally viewed that it should be free of regulation, that electronic commerce should be free, and that social platforms, that is, intermediaries, should not themselves be liable for content posted by third parties.Footnote 6 However, with the increasing influence of digital technology on everyday life, the concerns about human rights violations have also rapidly increased, making the issue of liability for the online infringement of human rights unavoidable. Modelling the liability of social platforms has become a pressing issue, along with emerging efforts to institutionalise the accountability of digital collective actors, which includes human–algorithmic association.Footnote 7 But to make social platforms liable, it is necessary to resolve the problem of the accountability of private actors for human rights violations traditionally immune to human rights challenges because social platforms are owned by private actors who also manufacture their contents, and coordinate and control them. To remedy the situation, different strategies are employed or offered.
First, the transnational nature of digital communication and unreadiness of states to step out of their traditional zone of control within their borders made room for social platforms to turn to self-regulation and develop what Gunther Teubner following David Sciulii calls ‘societal constitutionalism’.Footnote 8 At the expense of democracy, by accumulating powers traditionally seen as public, the private actors have become responsive to values of fundamental rights through private actions. The Facebook Oversight Board’s decision to uphold and partially revise Facebook’s decision to suspend former US President Donald Trump’s account indefinitely for the alleged influence of his posts on the violent attacks on Capitol Hill on 6 January 2021 clearly illustrates this trend.Footnote 9 What is more striking is that international law has encouraged such privatisation of public law, because it lacks legally binding instruments that would regulate the direct responsibility of non-state actors and profit-driven companies for human rights violations. Take, for example, the United Nations (UN) Guiding Principles on Business and Human Rights, whose greatest achievement suggests that corporate responsibility implies a negative obligation to respect human rights and a positive obligation to understand and mitigate the negative impacts on human rights with due diligence.Footnote 10 In the absence of state reactions, business actors, including social platforms, have become quasi-regulators.Footnote 11
Second, despite the limits of international law and the fact that the European Convention for the Protection of Human Rights and Fundamental Freedoms generally obliges only states, the European Court of Human Rights has attempted to limit the power of social platforms, by ruling that, in principle, social platforms, in the role of content providers, are liable for third-party content.Footnote 12 Although this approach has been subject to substantial criticism because it might encourage social platforms to behave like censors to avoid liability or arbitrarily remove content, it is worth noting that the Court has made an effort to model a liability regime for human rights violations in the online context through its doctrine of positive obligations.Footnote 13
Third, within the European Union (EU), a social platform’s liability was initially perceived as non-existent, providing that in a specific case it ‘has neither knowledge of nor control over the information which is transmitted or stored.’Footnote 14 In the opposite case, if it was established that the social platform possessed knowledge about the posted information and exercised control over it, the liability existed.Footnote 15 Today, the Court of Justice of the EU (CJEU) has become a significant regulator of the digital world, finding the balancing principle extremely helpful in disputes involving, on the one hand, the right of social platforms to conduct business, and on the other hand, the rights of individuals. Furthermore, the recently adopted Digital Service Act, the Digital Market Act, and AI Regulation represent additional EU attempts to force platforms to act more responsibly.Footnote 16
From a methodological perspective, although policymakers and judges have attempted to cope with the issue of intermediary liability for human rights violations within different fields, including data protection law, consumer law, privacy law, intellectual property law, and hate speech regulations,Footnote 17 on the theoretical level, the dominant way to frame this new phenomenon is to label it in constitutional terms. Starting from the fact that constitutionalism aims to limit government to protect individual rights, the emerging model of digital constitutionalism strives towards the same aim: to provide an understanding of how digital technology affects human rights, which are traditional tenants of constitutional law, and offer solutions for how to make private actors responsible for human rights infringements in the context of the digital world.Footnote 18
On this account, some scholars have offered the recognition of the new rights as the exit strategy. In addition to already available rights, the right to explanation (in the context of data processing), the right to accessibility, and the right to obtain a translation from the language of technology into the language of human beings are perceived as means to regulate the relationship among three main participants of the information society – platforms, states, and individuals.Footnote 19 Others examined whether the horizontal application of constitutional rights in private law may be a possible response to the unlimited powers of social platforms.Footnote 20 Several years ago, the Council of Europe’s Committee of Ministers acknowledged the significance of the horizontal effects strategy in regulating the responsibilities of online platforms and recommended to Member States to ensure the horizontal effects of constitutional rights in relations between private parties.Footnote 21
Considering that human rights violations in the digital sphere are of a constitutional quality, this chapter identifies the horizontal application of constitutional rights as a possible response to human rights challenges raised by the actions of social platforms. The traditional view in constitutional law is that constitutional rights are shields only against the state. However, in this chapter, I will start from the premise that the focus should not be on the state’s obligations but the individual’s rights. Following Joseph Raz, who claims that ‘rights precede obligations and therefore there is no closed list of obligations according to a certain law [but that] … changed circumstances can lead to the creation of new obligations according to an already existing law’,Footnote 22 I will presuppose that constitutional rights correlate not only with different duties but also with different duty-bearers concerning the fulfilment of duties. In light of this conclusion, it is evident that digital technology has made social platforms a prominent duty-bearer toward constitutional rights. I intend to make progress on the issue of the liability of social platforms for individual rights violations by suggesting the horizontal application of constitutional rights as an available strategy to remedy individual rights infringements in the online environment. The issue of whether constitutional rights should be restructured to protect from all intrusions of the digital world (e.g., the digital code) and not only from the activities of social platforms is outside this discussion.Footnote 23
This chapter is divided into five parts. After the introduction, in the second part, I will more closely explain the (non-)application of constitutional rights in offline private relations. In the third part, I will discuss different approaches to the emerging authority of constitutional rights online. The fourth part will advance understanding of how the horizontal application of constitutional rights in the online environment helps establish and maintain democratic control over digital technology. In the concluding part, I will summarize why an extension of constitutional rights in the digital sphere could be a driving strategy for protecting individual rights against the intrusive power of the algorithmic society.
4.2 Rights Talk in Private Law
4.2.1 The Meaning and Relevance of the Public/Private Law Distinction
A distinction between private and public law provides a ground for the systematisation of law.Footnote 24 It first appeared in sixteenth-century legal treatises, which, largely ignoring the contrast in Roman law between jus gentium and jus civile, made a sharp distinction between jus publicum and jus privatum.Footnote 25 Ever since, what belonged to the first and what was covered by the second category, however, remained a subject of vivid discussion.
Broadly speaking, private law traditionally encompasses the law of contracts, tort, property, business associations, commercial transactions, and related fields governing relations between individuals.Footnote 26 The state here appears as a mere arbiter of the rights and duties that exist between private parties and is not the party with the interest.Footnote 27
Figuring out the meaning of the term public law became more challenging. Thus, scholars have never had a complete control of the definition of public law. For example, public law was entirely omitted in Justinian’s civil law, but Hale and Blackstone’s civil law incorporated much of what the Romans would have called public law – including the rights and duties of the monarch, members of Parliament and other magistrates.Footnote 28 Issues of whether public law is an autonomous body founded on the autonomy of the political realm, whether it is isolated from morality, to which philosophy should we turn to specify its subject and tasks (e.g., to functionalist legal thought, legal positivism, Dworkinian legal interpretivism, or political theory), and which values are immanent within public law, have been subject to a modern passionate debate exemplified in the work of Loughlin, Craig, Harlow, and Cane.Footnote 29 At the highest level of abstraction, one may say that public law is inseparable from the government.Footnote 30 Following this argument, one can argue that constitutional law, criminal law, and administrative law are principal tenants of public law. Yet one should bear in mind that a lack of a clear definition of public law is also a result of still present differences among jurisdictions – despite emerging unifying trends, the tenants of public law are not the same in, for instance, France, the US, and England.Footnote 31
The issue of whether public law is fundamentally different from private law slowly loses its attraction since, in contemporary times, it is often unclear whether the relevant institute, value, or principle derives from public or private law. This is why Kelsen’s claim, that a distinction between private and public law is ‘useless as a common foundation for a general systematization of law’, is still valid.Footnote 32
There is much more to be said on the distinction between public and private law, but what helps approach the issue of the impact of constitutional rights on private parties is Kelsen’s view that traditionally, private law embraces norms governing relations between private parties, while public law embraces norms stipulating rights and duties between the state on the one hand, and private parties on the other.Footnote 33
4.2.2 The Riddle of Horizontality: From Natural Law to the Horizontal Enforcement of Constitutional Rights
Individual rights were first theoretically articulated in natural law theory without any differentiation between duty bearers responsible for their violations and the state obligation to protect individual rights regardless of who a perpetrator was:
The state of Nature has a law of Nature to govern it, which obliges everyone, and reason, which is that law, teaches all mankind who will but consult it, that being all equal and independent, no one ought to harm another in his life, health, liberty or possessions […]. The great and chief end, therefore, of men uniting into commonwealths, and putting themselves under government, is the preservation of their property […].Footnote 34
Thus, John Locke never considered natural rights to create obligations only for states, nor did he think a state’s duty to protect rights extended only to its intrusions. His vision found an expression in the first state documents stressing freedom and equality and acknowledging various duty holders vis-à-vis individual rights and a total state obligation to protect them from different intruders, not only from the state.Footnote 35
While the idea of natural rights mostly proved short, the constitutionalisation of individual rights in different forms and through different generations continued to endure after its first occurrence in the US Bill of Rights. However, over time, the focus on the state’s duties became the centre of constitutional protection as the state accumulated power and authority over its citizens.Footnote 36 As a result, constitutional rights, with fewer exceptions, extended only into the public law regime but not the regime of private governance. What followed on the theoretical level was the development of an animating idea that constitutional rights apply only vertically – in the relationship between a state and individuals – and not horizontally between private parties.
However, the fall of the Berlin Wall, the end of apartheid, and the return of many Latin American countries to democracy in the 1990s encouraged drafters of the new constitutions for post-communist, post-apartheid, and post-authoritarian societies, and constitutional law scholars, to rethink the nature and the purpose of the constitution, including the issue of its application in private law.Footnote 37 Furthermore, the phenomenon of globalisation, the increased activity of non-state actors in world conflicts, and the penetration of transnational corporations, non-government organisations, and digital platforms into the traditional public sector have imposed a critical theoretical question – who are the duty bearers to whom human rights (either constitutional or international) impose burdens and obligations?Footnote 38 Translated into this discussion, the starting position is that constitutional rights have become just as vulnerable to private actions as to the states, but unlike their constraint of the state in a Lockean sense, constitutional rights, in principle, do not constrain private actors.
Now, the academic discussion on whether constitutional rights should or should not produce effects in private law has brought to light three different positions regarding the applicability of constitutional rights in private law.
The first two positions are mutually exclusive. One, verticality, exposes the already elaborated traditional idea that constitutional rights protect only against the government and have no application in private law. Accordingly, they are judicially enforceable in public but not in private law. The main justification for insisting on verticality is the protection of individual autonomy in the private sphere, along with liberty or privacy, and the need for market efficiency.Footnote 39 A frequent argument is that autonomy in the private sphere should be isolated from any state action or control.Footnote 40 A constitution’s mandate is to secure a limited government and not to regulate private relations, which should be based on free individual choice.
The other position rests on the opposite, horizontal, approach: although defined and determined in constitutional (i.e., public) law, constitutional rights are directly applicable in private law, meaning that constitutional rights protect not only against government but also against private parties. Apart from insisting on the vulnerability of human rights in relation to any actor, either state or private, some authors also assert that there is no reason to avoid the constitutional regulation of autonomy for the purpose of its protection, as autonomy is already regulated by non-constitutional law.Footnote 41 Others indicate the function of a constitution, arguing that as a supreme law of the land it should apply to all equally. For example, Mattias Kumm’s total constitution claim derives from the premise that private law also involves political choices, and that political choices are subject to a constitutional rights review using proportionality analysis; this, in turn, suggests that decisions relating to private law should not be excluded from constitutional rights scrutiny.Footnote 42 Consequently, constitutional rights should be judicially enforceable in disputes between private parties.
The third position stands between the two extreme positions. It is grounded on an intermediate position, known as the indirect horizontal effect, and assumes that although constitutional rights apply directly only in public law against the government, they nevertheless indirectly apply, that is, produce effects, on private law.Footnote 43 While in the direct horizontal effect position, private actors are directly subjected to constitutional rights, in the indirect horizontal effects position, private laws are subjected to constitutional rights. Courts play a decisive role here as they must, in one way or another, take into account constitutional rights in deciding disputes between private parties.Footnote 44 What is indirect here is the fact that individuals are protected not directly by constitutional rights, but by the effects constitutional rights produce on private law.Footnote 45
There is a further wrinkle here. Constitutional rights can produce either strong or weak indirect horizontal effects on private law. A strong indirect horizontal effect assumes that all private laws are subject to constitutional rights and may be challenged in private litigations, meaning that individuals are fully (yet indirectly through private law) protected by constitutional rights. Contrary to this, a weak indirect horizontal effect means that private law is not subjected to constitutional rights, but that courts can take constitutional values exemplified in applicable constitutional rights into account when interpreting or developing private laws in litigation between private parties.Footnote 46 What stands behind the weak indirect horizontal effect is a claim that ‘the actions of private individuals can produce similar or identical effects or harms to those of governmental’.Footnote 47
4.2.3 How Horizontality Works Offline
The concept of the horizontal effects of constitutional rights is one of the basic coinages of modern constitutionalism. Although Ireland and South Africa are famously known for expanding constitutionalism in private law, horizontality is operational in some other jurisdictions, most notably in many Latin American countries, including Argentina, Bolivia, Chile, and Colombia,Footnote 48 and then in Malawi, Ghana,Footnote 49 and Slovenia.Footnote 50 As a rule, horizontality means different things in different jurisdictions.
The Irish Constitution itself contains an expressed commitment to direct horizontality: ‘The State guarantees in its laws to respect, and, as far as practicable, by its laws to defend and vindicate the personal rights of the citizen.’ Since 1965, the Irish courts have maintained that constitutional rights may have a direct horizontal effect and are not imposing obligations on the state alone.Footnote 51 The objections to the horizontal application of constitutional rights from liberal constitutional theory have played no role in judicial reasoning.Footnote 52
South Africa has a specific approach to the horizontality issue.Footnote 53 Apart from verticality, its 1996 Constitution endorses the direct horizontal application of constitutional rights in private law ‘taking into account the nature of the right and the nature of any duty imposed by the right’.Footnote 54 This is not the end of the story. The Constitution also authorises the indirect horizontal application of rights in disputes between private parties through the courts: ‘When interpreting any legislation, and when developing the common law or customary law, every court, tribunal or forum must promote the spirit, purport and objects of the Bill of Rights.’Footnote 55
An important footnote should be added here. Direct horizontality also works in France, but not as a constitutional matter. French legal culture does not recognise the horizontal effect of rights, and, in some way, it is incompatible with it because the state has never been perceived as a threat to rights but rather as the protector.Footnote 56 Nevertheless, rights do produce a direct effect in private law through the judicial application of the European Convention on Human Rights.Footnote 57
In Germany and Canada, constitutional rights do not have a direct but an indirect horizontal effect. In the world of horizontality, Germany is best known for the application of the Drittwirkung doctrine, meaning third-party effect, which has produced a substantial horizontal effect in jurisprudence. The doctrine was born in the jurisprudence of the German Federal Constitutional Court. By specifying its premises in the famous Lüth case decided in 1958, the Court ended a decade-long passionate discussion among German scholars and courts about the scope of the then newly adopted Basic Law (1949).Footnote 58 The Federal Labour Court took a leading role in the discussion, asserting that the constitutional rights, protected in the Basic Law, were directly applicable to relations between the employer and employees.Footnote 59 Concerned with the fact that ‘basic rights shall be binding for the legislative, executive and judicial powers’, the Federal Constitutional Court did not accept the standing of the Federal Labour Court. Yet, more importantly, it did not fully reject the idea that fundamental rights could have produced effects on the relations between private parties. The Court adopted what is now called the indirect horizontal effect model, in which constitutional rights were understood as legal codifications of objective general values immanent to the whole legal order, including private law:
This value system, which centres upon human dignity and the free unfolding of personality within the social community, must be looked upon as a fundamental constitutional decision affecting the entire legal system […] It naturally influences private law as well; no rule of private law may conflict with it, and all such rules must be construed by its spirit.Footnote 60
In short, although the Drittwirkung doctrine accepts that basic rights oblige only state organs, it nevertheless holds that: (a) all private law is directly subjected to constitutional rights and is invalid if it conflicts with constitutional rights, and (b) it is not on private actors to conform their actions with constitutional values, but rather, it is on judges, bound by the basic rights under Article 3 of the Basic Law, to consider constitutional values when interpreting private laws.Footnote 61
The German indirect horizontal effect model, with some variations, is also followed in Canada. Like Germany, Canada also supports verticality when it comes to the effects of the Charter of Fundamental Rights and Freedoms on legislation.Footnote 62 Yet the Supreme Court has distinguished between the constitutional rights and constitutional values embodied in the Charter, allowing constitutional values to influence the entire legal system, including private law, when ‘private litigant disputes fall to be decided at common law’.Footnote 63 This means that the principal role of the rights embodied in the Charter is to protect citizens against the government, but the courts may ‘apply and develop the principles of the common law in a manner consistent with the fundamental values enshrined in the Charter’.Footnote 64 Therefore, unlike in Germany, not all private law in Canada is directly subjected to constitutional review for its inconsistency with the Charter, but only common law, when the courts should check its consistency with Charter values (not rights). This solution directly derives from the separation of powers principle in the (British) common law tradition, under which the courts are allowed to develop common law in parallel with the Constitution.Footnote 65
Compared with Germany and Canada, the US is an outlier in the world of horizontality, although it adheres to the same starting position – the constitutional rights – as they oblige only state actors. Influenced by the strong liberal tradition, in which individual autonomy is highly cherished in all law, not only that the constitutional text refers explicitly to the obligation of states when conferring rights (No state shall…), but also that the US Supreme Court has established the controversial ‘state action doctrine’, which, arguably, precludes the influence of constitutional rights in private law. Therefore, in the original case from 1903 – The Civil Rights Cases – in which the doctrine was born, the US Supreme Court held that since they apply only to government actions, the Thirteenth and Fourteenth Amendments are not an appropriate basis for Congress to pass laws protecting African-Americans from discrimination.Footnote 66 The Court emphasised that constitutional rights shielded by the Fourteenth Amendment were so designed that it had only negative effects – they imposed duties of restraint only on federal or state governments whose duty was to protect them only from actions taken by the state. However, the Court made an exception regarding the Thirteenth Amendment, arguing that ‘Congress may probably pass laws directly enforcing its provisions, yet such legislative power extends only to the subject of slavery and its incidents […].’Footnote 67
The issue of who the state is for the state action doctrine has extended the reach of this doctrine to private parties providing that their conduct can be in different ways attributable to the state.Footnote 68 However, such situations are limited because the US Supreme Court has refused to find state action under many strands, apparently without clear criteria.Footnote 69 On the other hand, for those who are committed to debating the state action doctrine, a key concern is either which law is subject to constitutional scrutiny or whether courts, as state actors who enforce those laws, must be subject to constitutional rights scrutiny. Thus, focusing on the courts, the US Supreme Court itself ruled in Shelley v. Kraemer that it was unconstitutional for the courts to grant relief to enforce a racially restrictive covenant because this would constitute a state action under the Fourteenth Amendment.Footnote 70 Stephen Gardbaum even claims that the US adheres to the horizontal model when viewed through a comparative lens, as all law is fully and equally subject to constitutional rights scrutiny.Footnote 71 Nevertheless, the dominant position in the US is still that the Constitution limits the application of constitutional rights to public law. I hasten here to say that in cases involving constitutional rights infringements in a digital context the US Supreme Court has not shown any intention to change this position.
4.3 The Authority of Constitutional Rights Online: Emerging Trends and Resistance
Having documented the horizontal effects of constitutional rights offline, I now turn to present similar efforts to validate the horizontal effects of constitutional rights online. This section takes the freedom of expression and the right to privacy as the focus of attention.
4.3.1 Is the Internet a ‘New Free Marketplace of Ideas’ Immune to Constitutional Review?
There are several reasons why freedom of speech deserves special status in constitutional democracies. First, freedom of speech essentially contributes to social progress and the moral and intellectual development of individuals.Footnote 72 Second, without citizens being free to express, deliberate, and accept different ideas, there is no democratic government.Footnote 73 Third, freedom of speech is indispensable for establishing the truth: in the famous words of Justice Oliver Wendell Holmes, truth should not be regulated, but determined in the ‘marketplace of ideas’.Footnote 74
Probably the world’s best-known free speech clause is embodied in the US First Amendment. In case law, it is firmly established that the First Amendment does not permit the government to engage in a viewpoint-based regulation of speech without a compelling governmental interest, such as averting a clear and present danger of imminent violence.Footnote 75 Compared with the US, where the First Amendment provides almost an unlimited right to freedom of speech, the German Basic Law and the Canadian Charter, despite a presumption in favour of freedom of speech, offer mostly a qualified right, subject to limitations for different reasons. These different approaches have already been transplanted online. German and Canadian case law indicates that transplantation has followed the horizontality route. The American case shows that the US denied the horizontal applications of constitutional rights in the online context following its position in the offline world. Consider the following.
4.3.1.1 The American Approach: Ignoring Horizontality from LICRA to Gonzales and Taamneh
In 2000, the High Court in Paris famously ruled against Yahoo. The dispute began when two human rights organizations (La Ligue Internationale Contre Le Racisme Et l’Antisemitisme (LICRA) and L’Union Des Etudiants Juifs De France) sued Yahoo in France for allowing its users to offer Nazi-related items for sale on Yahoo.com, as the sale, exchange, or display of Nazi-related materials or Third Reich memorabilia, represented a hate crime outlawed by the French Penal Code.Footnote 76 The French Court ruled that Yahoo’s auction site violated the Penal Code and bluntly ordered Yahoo to preclude access to the auction site and other sites displaying Nazi-related material for French citizens and warn its users to refrain from accessing content prohibited by French law to avoid legal sanctions.Footnote 77
In the US, however, the claim that everyone ought to have their rights protected against everyone, whether offline or online, is not appealing. The argument resurfaces in the LICRA case. Because the dispute also involved jurisdictional issues, the US District Court for the Northern District of California was asked by Yahoo to intervene and declare that the French Court’s decision in LICRA was neither recognisable nor enforceable in the US.Footnote 78 The US Court specified that the lawsuit aimed to determine ‘whether a United States court may enforce the French order without running afoul of the First Amendment’.Footnote 79 It issued a declaratory judgment and ruled, inter alia, that ‘Yahoo has shown that the French order is valid under the laws of France, that it may be enforced with retroactive penalties, and that the ongoing possibility of its enforcement in the United States chills Yahoo’s First Amendment rights.’Footnote 80 In the view of the US Court, the French Court’s demand that Yahoo ‘take all necessary measures to dissuade and render impossible any access via Yahoo.com to the Nazi artifact auction service and any other site’ was too general and imprecise and amounted to censorship of protected speech.Footnote 81 The decision follows the US Supreme Court’s finding that a law may violate the First Amendment if it is so ‘overly broad’ that it infringes protected and unprotected speech.
At this point, the basic considerations should be clear. In its landmark ruling on the online freedom of expression in Reno v. ACLU, the Supreme Court declared unconstitutional two provisions of the Communications Decency Act that criminalised ‘obscene or indecent’ speech transmitted to children and the delivery of ‘patently offensive’ information to children.Footnote 82 As a matter of fact, social platforms are exempted from liability for material posted by someone else on their sites, regardless of whether the posts violate the right to free speech. In such cases, Section 230 of the US Communications Decency Act exempts internet platforms from liability expressly providing that they would not be treated as ‘publishers or speakers’.Footnote 83 Moreover, the established case law testifies that the state action doctrine does not apply to social platforms. The US federal courts have repeatedly rejected the notion that private corporations providing services via the internet are a public forum for the purposes of the First Amendment. Take, for example, the cases of Dipp-Paz v. Facebook and the Federal Agency of News v. Facebook.
In the Dipp-Paz case, the plaintiff asserted that Facebook violated his constitutional right to free speech by blocking his account.Footnote 84 However, the Court dismissed the claim finding that the plaintiff did not show that Facebook ‘acted under the colour of a state “statute, ordinance, regulation, custom or usage”’.Footnote 85 The Court denied that Facebook is a public forum to which the First Amendment requirements applied, stressing that ‘Facebook is a private corporation, and Plaintiff does not allege any facts suggesting that Facebook’s actions are attributable to the state.’Footnote 86 The District Court for the Northern District of California followed the same reasoning in the Federal Agency of News (FAN) case, when Facebook blocked and removed the account of the Russian agency (FAN) allegedly involved in the US presidential elections in 2016.Footnote 87 The Court ruled that Facebook did not operate as a public forum and that its actions did not amount to state action under the public function test on the ground that Facebook was not a wilful participant in joint action with the government, nor did it conspire with the government to violate any constitutional rights.Footnote 88
Some hope for a change was raised when recently the US Supreme Court was asked in Google v. Gonzales and Twitter v. Taamneh to shift the foundations of internet law by narrowing or revoking the protection that Section 230 secures for online platforms.Footnote 89 Both cases were initiated by the families of victims of ISIS terrorist attacks who alleged that Twitter and Google-owned YouTube helped the ISIS group carry out the attacks. Both opponents and proponents of Section 230 anxiously awaited the decisions. Eventually, the Supreme Court said nothing about Section 230 in either ruling. In Twitter v. Taamneh, it ruled exclusively on the grounds of the Justice Against Sponsors of Terrorism Act. After considering whether the defendant’s conduct constituted aiding and abetting by knowingly providing substantial assistance, the Supreme Court found that the plaintiffs failed to establish their case. In the Gonzales case, the Court openly declined to address the application of Section 230 and, by unanimous vote, returned the case to the lower Court to rehear it in light of its decision in Twitter v. Taamneh, implying again that there was no need for Section 230 to be addressed.
The Supreme Court’s approach to these decisions could be read differently. Thus, one may claim that the Court purposely followed the judicial minimalism strategy to allow the voters and Congress to decide whether they want the internet to change. In constitutional cases of high complexity, which divide people on moral or other grounds, a minimalist path makes sense because democracy then urges the legislature to decide.Footnote 90 Any limitation of the freedom of speech, particularly in the context of new or changing circumstances, is such a case in the US. The narrow rulings could also suggest that nine justices appeared to postpone the ruling on Section 230 as ‘other cases presenting different allegations and different records may lead to different conclusions’, as Justice Jackson observed concurring in Twitter v. Taamneh. This may be so, particularly if one takes into account yet another possible reading. Strictly speaking, in Gonzales and Taamneh, the Supreme Court did not say online platforms were protected under Section 230, but rather it found no direct link between the terrorist attacks and online posts and videos. Notwithstanding which of these readings holds promise from a constitutional rights perspective, the Supreme Court’s decision to avoid considering Section 230 alone represents a victory for online platforms, at least for the time being.
4.3.1.2 German and Canadian Approach: Horizontality Matters in Online Speech
Because social networks took over a significant portion of the public sphere, the constitutional dimension of their responsibility attracted profound attention among scholars in Germany. An initiative that called for intermediary responsibility under the same standards as the state was not taken seriously, but a discussion on the horizontal effects of the freedom of speech and the connected right of the platform to delete content gained significance.Footnote 91
The approach insisting on the transplantation of the Drittwirkung doctrine in the digital sphere obtained judicial recognition. In its decision delivered in 2021, the German Federal Court of Justice, in a case involving hate speech online, took a more balanced approach than the French court in LICRA and solved the issue on the grounds of the indirect horizontal effect of constitutional rights.Footnote 92
The Court was faced with two cases involving the Facebook decision to delete posts and partially block users’ accounts with the explanation that hostile remarks about migrants amounted to hate speech. Unlike the French Court, which in LICRA paid no attention to the interests of Yahoo, the German Federal Court of Justice was more cautious. It balanced two constitutionally protected and conflicting rights: users’ freedom of expression, protected under Article 5, and Facebook’s occupational freedom, covered in Article 12 of the German Basic Law.Footnote 93 The Court found that, based on its right to occupational freedom, Facebook, in principle, was entitled to require users to respect specific communication standards and to block users’ accounts responsible for possible breaches. At the same time, it emphasised that Facebook’s right to occupational freedom is not unlimited. Facebook’s terms of business, including standards for deleting posts or blocking users for a breach of standards, must, following Article 307 of the German Civil Code (the requirement for reasonable business terms), take into account the involved fundamental rights, as in this case, the freedom of expression.Footnote 94 Therefore, the Court ruled that Facebook’s business terms related to deleting users’ posts and blocking accounts in case of violation were invalid. Before blocking them or deleting the content that amounted to hate speech, Facebook should have consulted the affected users, informed them about the deletion, and made possible redress opportunities after the deletion.Footnote 95
Eventually, the German Federal Court of Justice did not rule that Facebook, as a private company, had a constitutional obligation to respect freedom of expression. Rather, it used the Drittwirkung doctrine to order Facebook how to delete posts and block users’ accounts for posting content with hate speech implications. The portion of the Court’s decision requesting information obligations and complaint mechanisms mirrors the same solution embedded in the German Network Enforcement Act (NetzDG).Footnote 96 It also corresponds to the solutions proposed in the EU Digital Services Act, which requests online platforms to comply with obligations related to transparency, information obligations, and complaint mechanisms in relation to the removal of illegal content and the protection of users’ fundamental rights online.Footnote 97
Now, the situation in Canada is intriguing. The 2020 United States–Mexico–Canada Agreement limits the civil liability of online platforms for third-party content and does not treat them as content providers.Footnote 98 However, even since 2005, in the defamation context, the Canadian courts have ruled that online platforms could be liable for defamatory comments posted by third parties under certain circumstances.Footnote 99 Yet a breakthrough has been recently announced thanks to the availability of the horizontality doctrine. I will explain this in more detail by tracing the decision in the ongoing lawsuit in Cool World Technologies Inc. v. Twitter Inc.
Who governs online content in Canada is a major issue in this case initiated against Twitter for rejecting to promote posts relating to a Canadian documentary film.Footnote 100 The applicant (the publicity firm Cool World Technologies) alleged that Twitter wrongfully refused to sell advertising space on the Twitter social media platform, which ended in the violation of the applicant’s freedom of speech.Footnote 101 The applicant relied on the influence of constitutional values related to the freedom of speech in Canadian contract law to support their claim. Because Twitter in Canada has such a political significance as to represent the ‘town hall’, the applicant asserted that the public policy of Canada, ‘informed by Charter values related to freedom of expression, preclude Twitter from enforcing contract terms to exclude high value, non-harmful speech from its self-proclaimed town hall’.Footnote 102
On the other hand, Twitter claimed that it had absolute and unfettered discretion to refuse any advertising posts and that all of Twitter’s user accounts had no nature of the governing contract.Footnote 103 Besides, Twitter urged that the applicant was not entitled to allege the breach of Charter values, reminding that no right for one private party to sue another private party for breach of the Charter or breach of analogous Charter values existed.Footnote 104
At the preliminary stage, the Court allowed the lawsuit to proceed, finding that the applicant could base their case on the effects that the constitutional guarantees of freedom of speech produce in contract law because of Twitter’s central role in the public life of Canada.Footnote 105 Whether Twitter’s unlimited power to control the content on its platforms in Canada will remain untouched by the end of the judicial proceedings remains to be seen. However, the indirect horizontality effect of constitutional values has opened the door for the judicial protection of freedom of speech on social platforms in Canada.
4.3.2 Horizontality in Service of Privacy-Related Rights in Online Contexts
Privacy-related rights are also frequently exposed to gross infringements in the digital environment. For the time being, all we know about the horizontal effect of privacy and privacy-related rights in the digital context mostly comes from the EU and the pioneering practice of the CJEU.
It has been quite a while since the constitutional nature of EU primary law and the constitutional value of rights in its legal order were emphasised.Footnote 106 The horizontal effects of certain equality-related rights were announced already in the Rome Treaty adopted in 1956. This came to the surface when the CJEU in 1976 ruled that Article 119 of the Rome Treaty, ensuring the principle of equal pay for male and female workers for work of equal value, obliged not only the Member States to whom the provision was directed but also private employers.Footnote 107
In EU law, considerable attention has also been given to the horizontal effects of the EU Charter of Fundamental Rights. Although it has the same legal effect as the EU Treaties, the horizontal effect of its provisions has provoked much debate. The CJEU implicitly confirmed that, when applied within the scope of EU law, the Charter could create obligations for private parties if its provisions grant a legal right for an individual and not just a principle.Footnote 108 This finding seems reasonable, considering that the Charter has the status of EU primary law. The ultimate confirmation came in 2018 when, in four decisions, the CJEU established the direct horizontal effect of several Charter rights in disputes between private parties, specifically the right to non-discrimination, certain rights related to fair and just working conditions, and the right to an effective remedy and a fair trial.Footnote 109
However, even before its 2018 revolutionary offline case law, the CJEU opened the door to the indirect horizontal application of fundamental rights online, in particular, to privacy-related rights, embodied in Articles 7 and 8 of the Charter. In the Google Spain case, it articulated the right to be forgotten, relevant in the framework of the right to data protection and freedom of expression and information.Footnote 110 The case involved the interpretation of the Data Protection Directive in the dispute between Google and the Spanish data protection agency. It concerned the removal (delisting) of personal data available online. Among several questions addressed to the CJEU, important for this discussion is the question of whether, under the Directive, an individual who does not wish to make personal data available to internet users has the right to address a search engine directly and ask it to delist the personal information published on third parties’ web pages.Footnote 111
The CJEU resolved this question by famously concluding that if the activity of a search engine significantly affects the fundamental rights to privacy and the protection of personal data, the operator of the search engine ‘must ensure […] that the guarantees laid down by the directive may have full effect and that effective and complete protection of data subjects, in particular of their right to privacy [….]’.Footnote 112 Although the said Directive did not explicitly create the right for an individual to request the data processor to remove their personal data, the Court stressed that the Directive had to be interpreted as if it included such a right because of the effects the rights to privacy and the protection of private data, guaranteed in Articles 7 and 8 of the Charter, produced on the Directive.Footnote 113 Consequently, the CJEU concluded that, ‘the data subject may, in the light of his fundamental rights under Articles 7 and 8 of the Charter, request that the information in question no longer be made available to the general public by its inclusion in such a list of results’.Footnote 114
Although certifying that the Charter produced an indirect horizontal effect on EU secondary legislation, the CJEU nevertheless recognised that under the given circumstances, the legitimate general interest of the public in accessing information also existed. Balancing thus became necessary: ‘A fair balance should be sought in particular between that interest [the legitimate interest of the public in accessing information] and the data subject’s fundamental rights under Articles 7 and 8 of the Charter.’Footnote 115 The burden of the balance was put on the search engines, which, according to the CJEU, qualified as personal data controllers within the meaning of the Data Protection Directive and ‘within its responsibilities, powers, and capabilities’.Footnote 116 According to this view, the CJEU acknowledged that these rights override, as a rule, the economic interest of the search engine operator and the general public’s interest in having access to that information.Footnote 117
Therefore, the ultimate outcome of the case carries the particular interest of the individual to have personal data delisted, the general interest of the public in accessing information, as well as the obligation of the search engine (e.g., Google), to balance the relevant rights when assessing users’ requests to delist personal data from search results. The broader point here is that the Google Spain case exemplifies how consequential the indirect horizontal application of the Charter can be in practice, as the Court made both the Charter and the said Directive applicable in disputes between private parties.Footnote 118 Moreover, via its readiness to vest the Charter’s privacy-related rights with horizontal effects, the CJEU reaffirmed another landmark decision on digital privacy in Schrems I – that the Data Protection Directive, since it regulates the processing of personal data and is liable to infringe fundamental freedoms, in particular, the right to respect private life, must always be interpreted in light of the Charter’s rights.Footnote 119
4.4 How the Horizontality Doctrine Helps Prevent Digital Threats to Democracy
The examples from German, Canadian, and the CJEU constitutional jurisprudence testify that the horizontal application of constitutional rights can accommodate both the concerns of those who object to any limitations to what is termed ‘internet governance’ and of those who insist on protecting rights online in the same manner as they are protected offline. Horizontality operating in the online world can also increase the democratic legitimacy of the online world. I will now turn my attention to this conclusion.
There is good reason to believe in the potential of the internet to upgrade democracy. Online communications and deliberations could help in developing an ideal, an internet-facilitated public sphere, with free discourse that could legitimise democratic government in Habermas’s sense.Footnote 120 Illuminating such potentials, the UN Human Rights Council concluded that ‘facilitating access to the internet for all individuals, with as little restriction to online content as possible, should be a priority for all States’.Footnote 121
However, there is also good reason for concerns about digital threats to democracy, ranging from weakening election and referendum integrity (e.g., in the 2016 US elections and the Brexit referendum campaign) to disinformation or hate speech in the online public sphere and interference with opinion formation in social networks, especially when they produce extremism.Footnote 122 In the digital world, the allegations that democracy implies only the majority rule and that a democratic system should be highly responsive to popular will are more easily sold to the public than in the offline world. Consequently, social practices that form preferences may question the legitimacy of decision-making processes. In particular, as Stephen Holmes suggests, the individuals whose activities are left without constraints exercise more significant influence than those responsible for making decisions.Footnote 123 In the presence of the unconstrained majority, the garden-variety examples either from the offline or online worlds testify that individual rights are the first to suffer in these circumstances.
Now, what a democratic constitution tends to achieve is to minimise the tension between democracy and individual rights, as people are prone to overstate this tension.Footnote 124 Moreover, apart from the usual argument that human rights are undemocratic, some took comfort from the observation that individual rights could be reconciled with democracy only if perceived as serving majorities.Footnote 125 Yet, on this account, Cass Sunstein essentially denies that democracy is an antagonist to rights. On the contrary, a democratic constitution, he claims, protects rights and thus constrains ‘what majorities can do to individuals or groups’.Footnote 126
Following Sunstein, it is not hard to conclude that no other strategy to tame the power of the online platforms brings the online world closer to democracy than the radiating effect of constitutional rights on the internet. The explanation of why democratic control matters here is almost self-evident. On the one hand, in a functional constitutional democracy, human rights are subjected to effective protection, while sanctions for their violations are pre-conditioned by the government’s democratic legitimacy, the rule of law, transparency requirements, and accountability under constitutional rules. On the other hand, the idea that the self-binding could be a strategy for the digital world was not comprehended initially. Those who maintained that some regulations were needed advocated the creation of a new system with no clear parallel in the offline world.Footnote 127 When internet governance emerged, it was established exclusively within the venue of private powers.
Thus, it turned out that those who make laws in the digital environment (a) order who can participate and who cannot in online communication, (b) design protocols and procedures in the case of digital rules’ violations, and (c) are private entities, including online platforms and their self-regulating bodies. How they make choices related, for example, to freedom of speech, the protection of privacy, or the collection of personal data, even when they do it under formally observed human rights law, such as Facebook’s Oversight Board, is left without democratic oversight. In other words, the power of the online platforms and their self-regulating bodies to design rules and control cyberspace compared with their accountability is underproportioned if not non-existent. The digital space’s private order suffers from a democratic deficit, which, interpreted by Haggart and Keller, exists since ‘private companies make the choices that set norms and directly influence the behavior of billions of users’.Footnote 128
Adjusting the constitutional system to the horizontal effects of constitutional rights is a reactive strategy to what happens online, but it legitimises the rules affecting individual rights and delivers results grounded on the citizens’ perceptions of what ‘the correct outcome is’ whenever rights should be balanced against general interests. In principle, when applied horizontally, the right to privacy or freedom of speech does not automatically prevail over the right of the online platform to conduct business but requires balancing, a process legitimised in the constitutional discourse whenever a court is asked to set aside a regulation of whatever kind, on the grounds of its non-compatibility with some constitutionally protected right.Footnote 129 On balance, it seems that under the horizontal model, no one loses; only democracy gains.
4.5 Conclusions
I have arrived at the end of a long trail of arguments offered to show why constitutional rights should be extended into the regime of internet governance, in particular to social platforms.
Social platforms have proven track records as to their capacity to pose harm to constitutional rights, which, according to the European Court of Human Rights, is even greater than that posed by the press.Footnote 130 Although the view that constitutional rights protect all private persons but oblige only the state has been abandoned in some jurisdictions under the doctrine of the horizontal effects of constitutional rights, and although this doctrine is a ready-made vehicle to make social platforms responsible for the intrusions upon constitutional rights, the claim that everyone ought to have his or her rights protected against everyone is still disputable. To remind the reader, the division between public and private law still dominates in constitutional systems across the globe: it mirrors the position that individual rights impose obligations only on the state and not private actors. Therefore, rights do not regulate relations between private parties whose autonomy should remain free from the compulsory regime created by constitutions. However, knowing that the state is not the only bearer of political and economic power and that individual rights are also threatened by private actors, including those operating in the digital world, the rhetoric must be changed. This step does not require the recognition of new rights but the recognition of new duty holders in relation to existing rights, such as social platforms. The examples from Germany, Canada, and the EU, jurisdictions traditionally open to the horizontal enforcement of constitutional rights, illustrate its promising potential to remedy human rights abuses that happen online.
Open access