2.1.1 Characteristics

Emulated Guardian is the overarching term used here for bodies like ODSs or the Oversight Board. The term points to both bodies’ two key characteristics. First, they are presumed to act as guardians of individual rights, in a similar way that courts or other adjudicators became guardians that protect individuals from overweening public power. I use the term guardian to reminisce the famous “who guards the guardians” issue in public law.Footnote ¹ Second, they are novel bodies whose structural features reflect or emulate ideas regarding judicial protection, individual rights, adjudication, and judicial review.

Emulation describes how organizational actors, such as the EU or Meta, build new policies and formal structures by imitating older institutionalized policies from other contexts. Emulating existing institutions has two advantages: first, emulating something is easier than entirely reinventing the wheel.Footnote ² Second, because institutionalized policies and formal structures are generally (that is, publicly) considered legitimate, emulating these to address novel problems maps the perception of legitimacy from the old context onto the new institution.Footnote ³ Simply put, emulating adjudicatory control over public – and especially administrative – power promises a smooth path toward controlling social media platforms and enables those social media platforms to present themselves as (more) legitimate.

Viewed through a legal lens, emulation is an unwieldy category, because actors like ODSs or the Oversight Board do not fit traditional conceptual frames. By dint of their emulative natures, ODSs and the Oversight Board reflect diverse aspects of various adjudicatory role models without, however, becoming the “original” – either in concept or in practice. Recent conceptual takes on Emulated Guardians range from supreme court,Footnote ⁴ to sub-courts,Footnote ⁵ de facto courts,Footnote ⁶ to human rights tribunals,Footnote ⁷ to (as yet unspecified) transnational adjudicators.Footnote ⁸ But, in my view, superimposing traditional concepts onto novel phenomena might risk “foreclos[ing] rather than foreground[ing] debates” about the current reshuffle of power relations at play.Footnote ⁹ In turn, to borrow Martin Shapiro’s words, “[i]f we turn to the work of those persons and institutions to which we normally award the titles judges and courts, we shall see that in reality they are simply at one end of a spectrum rather than constituting an absolutely distinct entity.”Footnote ¹⁰

On the contrary, analogizing amorphous and ambiguous emulators like ODSs or the Oversight Board to existing – and often patently legitimizing – ideas like “supreme” or “quasi-courts,” attaches public legitimacy and an assumed certainty to phenomena (yet) undeserving of these epithets.Footnote ¹¹ Simultaneously, as outlined in the introduction, pressing Emulated Guardians in schematic templates of public or private oversimplifies reality. For example, ODSs are formally private bodies. Yet, they only exist because the EU envisaged them in the DSA and national administrators license them. The EU’s “vision” for ODSs in turn, albeit formally emerging in legislative procedures, was shaped at least in parts by lobbying and backchanneling during legislation. Tucking phenomena into somewhat undercomplex conceptual boxes like “public” and the “private” is important to highlight indeed existing and important differences, yet the reality in those boxes seldom manifests as purely as dichotomous labels would suggest.

Therefore, I assume a deferential approach toward the criteria of adjudication just outlined. I investigate the diverse and at times convoluted practices in the case studies through the prism of individual rights adjudication as their paragon, point of reference, and model – but not as a strict deductive category.Footnote ¹² The practices of the Oversight Board and the out-of-court dispute settlement bodies appear to be heavily inspired by individual rights adjudication. Yet this observation does not necessarily tick all the boxes.Footnote ¹³ For example, the out-of-court dispute settlement bodies cannot bindingly decide cases and Meta retained some influence over the Board’s composition and docket, thereby undermining its independence.Footnote ¹⁴ Nonetheless, both bodies mirror many functional aspects of adjudicators and performatively play with their (limited) reproduction of adjudicative features like individual appeals, impartiality, norm-based decision-making, and juridical “habitus.”Footnote ¹⁵

The key trait of Emulated Guardians is, therefore, not that they are indeed courts or human rights tribunals but that they remind us of adjudicators without being “classic” specimens of the species. Resemblance without identity, one could say. Thus, focusing on the aspect of emulation acknowledges Emulated Guardians’ reflexive status as operating in a space between law and self-regulation, public expectation, self-serving interest, institutionalized and culturally dependent beliefs about controlling power structures, and the real practice of controlling power structures.

2.1.2 Performativity

If Emulated Guardians’ characteristic is that they oscillate between positive law and self-regulation, public imagination and actual authority, ceremony and actual control, they are quintessentially performative. In other words, Emulated Guardians’ practice, symbolism, and perception reflect changes in common and legal discourse and, reflexively, may change the very discourse itself. The idea is simple. Emulated Guardians can – and doFootnote ¹⁶ – bring with them a rich arsenal of symbols, practices, and argument-based resources that might channel some of the public legitimacy attached to classic adjudicatory institutions.Footnote ¹⁷ They appear – and are designed to appearFootnote ¹⁸ – as adjudicative institutions. However, compared to traditional individual rights adjudication, Emulated Guardians suffer from in-built design flaws. For example, the Oversight Board remains financially dependent on Meta, and neither ODSs nor the Oversight Board can issue binding decisions at scale.Footnote ¹⁹

Yet if their practice is perceived as sufficiently adjudicative, they may be able to incrementally build their own public legitimacy and authority vis-à-vis the power structure they shall control. Through practice and narration, Emulated Guardians may dress themselves in the publicly legitimizing guise of adjudicatory appearance. Because adjudication is such a powerful and revered societal institution Emulated Guardians may construct their own authority, outgrow inbuilt design flaws, and yield larger cultural changes toward more meaningful accountability. That would be public legitimation by emulation or, from a cynical point of view, self-legitimation by appropriation.

In turn, Emulated Guardians’ practice, which is related to but not quite adjudication in a classic sense, might reflexively alter the shared understanding of how individual rights adjudication works in the digital sphere or, perhaps, in the whole realm of post-state governance. We can observe similar performative reconstructions of crucial social concepts in many aspects of life.

Take the example of privacy. Changing societal perspectives on privacy and its enforceability offer a cautionary tale in a closely related field.Footnote ²⁰ Only two decades ago, much of the hyper-personalized data mining of Amazon, Google, and Meta would have arguably been viewed as an outrageous breach of individual privacy. In her comprehensive genealogy of what she calls “surveillance capitalism” in Silicon Valley during the 1990s and early 2000s, Shoshana Zuboff points out Google’s pioneering role in privacy-eroding “datafication”:

Yet by virtue of practice and rhetoricFootnote ²² (especially through imaginaries of endless connectivity, possibility and, ironically, individual freedom) our conceptions of privacy, or at least of privacy’s enforceability, have fundamentally changed.Footnote ²³ Quantitative empirical research found that “previous privacy violation experiences” had only “small effects on … subsequent online privacy concerns and no substantial effects on online information disclosure.”Footnote ²⁴ In other words, past experiences of eroding privacy do not raise a great deal of concern and most of the time do not elicit behavioral consequences, like withdrawing from social media platforms. Thus, the shared understanding of privacy changed.

Qualitative empirical findings corroborate these quantitative results. For example, after a privacy breach of eighty-seven million North American Facebook accounts by Cambridge Analytica in 2018, many users around the globe (especially younger ones) appear to have changed their perception of what Facebook could legitimately do with its – and by extension, their – data.Footnote ²⁵ Instead of anger or discomfort, which could have changed their usage of social media, some people began to modify their perception of privacy. As Hagar Afriat et al. argue:

Such changes of overall perception induced by changing practice and language can be (broadly) described as performative.Footnote ²⁷ The iterative repetition of concepts, practices, and symbols in different times and places attaches a general meaning to these concepts, practices, and symbols, but over time alters the general meaning of these concepts and the significance of practices and symbols pertaining to them.Footnote ²⁸

The formative effects of practice and language on reality, power, politics, and, therefore, law should not be misunderstood as a merely constructionist or postmodern quip.Footnote ²⁹ History is full of examples, ranging from the extreme to the mundane. A vivid example of the former were the “propagandistes” in post-Revolutionary France who sought to revamp the entire French language by fixating revolutionary spirit in everyday parlance to enable (but also to guardrail) citizens to speak (and think) only in revolutionary terms.Footnote ³⁰ Many examples of performative change and incremental shifts in meaning are less flashy and abrupt than the one just mentioned. Arguably, those changes that really stick are not paternalistically declared, but gradually normalized.

2.2.1 Individual Rights

Individual rights adjudication refers to rights-based, adjudicative ex-post control over power structures. If, for example, a police officer exceeds their power against a protestor, the latter may initiate individual rights adjudication, claiming civil rights violation and the likes. Individual rights adjudication rests on two pillars. These are, on one side, power mediators in the hands of individuals, or what I call broadly individual rights.Footnote ³² In the police example given above, the relevant individual rights could be those to bodily integrity, free expression, and freedom of assembly. On the other side, individual rights adjudication relies on adjudication as a mechanism to effectuate these power mediators. The adjudicator would be, in all likelihood, some kind of court.

The normative basis of individual rights adjudication are individual rights. The term “individual rights” is a functional concept. Functionally, individual rights can be viewed as normative practices that enable individuals (or other rights holders) to make normative claims to limit the exercise of power on the part of a dominant power structure.Footnote ³³ Individual rights can be human, constitutional, civil, or fundamental rights (depending on one’s jurisdiction) and other more specialized or derivative rights. Individual rights can be substantive (e.g., the right to freedom of expression) or procedural (e.g., the right to judicial protection). Individual rights need not follow a specific terminology or form, nor must they share a specific normative or philosophical underpinning.Footnote ³⁴ The key idea behind individual rights, which is the only relevant aspect at this point, is that they provide the basis for normative claims formulated by individuals to structure their relationship with a powerful actor. In this sense, rights are solely a functional category. They are “usefully understood as power mediators, as normative [practices] that materially weak actors can invoke to alter the power relationship between themselves and materially preponderant actors or institutions (empires, sovereign states, etc.).”Footnote ³⁵ That could entail, for example, claims that a powerful actor must not curtail personal liberties or take property without due process, cannot discriminate based on gender or race, or is prohibited from constraining intra-European trade for protectionist reasons.

Equalizing these power relations by exercising individual rights inevitably individualizes said relationships because it shifts the focus from underlying, structural factors to the exercise of power vis-à-vis individual rights holders. This aspect of individualization simultaneously constitutes individual rights’ main advantage and impediment. On the one hand, individual rights grant far-reaching agency to individuals, allowing them to challenge the exercise of power and thereby to act as agents for the common good.Footnote ³⁶ On the other hand, many rights violations occur arguably not as individual transgressions but reflect more deeply underlying issues relating to embedded structures, habits, procedures, and policies. Salvaging those structural issues requires proactive reforms that regularly go beyond vindication of individual legal guarantees.

2.2.2 Enforcement and Adjudication

Individual rights require an enforcement mechanism. Only if enforced, individual rights function as “power mediators.” Various enforcement mechanisms exist. Every individual can verbalize a claim, directly or implicitly referencing that they believe a right corroborates their normative position. Sometimes, the power structure might respond to such voice and voluntarily subject itself to cogently formulated arguments.Footnote ³⁷ However, that is not always the case. Therefore, many “guardians” of (public) power exist. We are familiar with various accountability mechanisms, procedures, and entities that hold to account large power structures, especially entities embodying public power. We may distinguish ex-ante influence exercised via participatory mechanismsFootnote ³⁸ from continuous oversight and scrutiny by elected representatives,Footnote ³⁹ auditors,Footnote ⁴⁰ ombudspersons,Footnote ⁴¹ the public,Footnote ⁴² or the press. Further, many administrative organizations feature individually accessible, internal though not independent review mechanisms.

However, one of the oldest, most institutionalized and revered practices for enforcing individual rights vis-à-vis administrative power is adjudication. Adjudication is characterized here as independent, norm-based dispute settlement between an individual and a power structure. “Adjudicator” refers to the independent entity that resolves the dispute and thereby exerts control over the power structure.

Depending on time and context, adjudication is formalized and institutionalized in various shapes and with diverse labels. The most prominent term for adjudicators is the “court” label.Footnote ⁴³ But there are other names as well. Adjudicators controlling administrative power might be called (administrative) tribunals,Footnote ⁴⁴ administrative courts,Footnote ⁴⁵ commissions,Footnote ⁴⁶ boards,Footnote ⁴⁷ or committees.Footnote ⁴⁸ Adjudication as a control mechanism of rule enforcement and rulemaking exists on the international level,Footnote ⁴⁹ in the EU,Footnote ⁵⁰ and within any rule of law system. Generally speaking, adjudication occupies a special place on the cusp of our sociocultural imaginary of taming the sovereign, advancing the rule of law, and making the putatively “liberal” political project a reality. Over time, adjudicators have also devised strategies for moving beyond individualized vindication of rights and aim at ameliorating those structural governance problems that give rise to individual rights violations in the first place.Footnote ⁵¹ Yet remedying structural governance flaws inductively by adjudicating individual rights remains complicated for several reasons – accessibility, adjudicative expertise when it comes to administrative practices, and advisable adjudicative restraint due to the notion of tripartite government.

Demands for, and the affirming of, the practice of adjudicatory control of administrative power are not limited to traditional administrative contexts. In modified form, adjudication that seeks to protect individuals from dominant power structures exists in various governance frameworks. For example, the UN Guiding Principles on Business and Human Rights oblige private power structures to be subject to public judicial and administrative review mechanisms and encourage businesses to make use, if necessary, of independent review mechanisms. Referring to the entire range of administrative or administration-likeFootnote ⁵² actors, Kingsbury, Stewart, and Krisch noted,

In sum, practice and theory have yielded idealized and widely diffused concepts of adjudication that build on and structure adjudicators, which again are much more diverse than purported by idealized notions. Further, many scholars, practitioners, and discourses identify adjudicative control of power as something we aspire to and desire.

2.2.3 Disputes and Power Relations

The ignition spark of adjudication is a dispute or, in other words, the desire to control an adverse, potentially dominant power structure.Footnote ⁵⁴ Regularly, the dispute concerns some form of prior exchange, act, or communication. Further, an individual’s desire to “control” a potentially adverse, dominant power structure. If there is no such desire, then there is no dispute, and consequently no need for adjudication. The disputes discussed in this book arise as a result of power imbalances, namely between individuals and social media platforms’ content moderation policies.Footnote ⁵⁵ The disputes concern how social media platforms unilaterally regulate and police the communication of their users. As this chapter explains in greater detail in Section 2.3, people often have little possibility to exit the platforms due to network effects and generally have no say in the rules applied to them.Footnote ⁵⁶

2.2.4 Impartiality and Independence

To settle disagreements about the exercise of power that cannot be resolved directly between the individual and the power structure, the individual might turn to a third party.Footnote ⁵⁷ That third party, the adjudicator, provides the forum for effectuating individual rights as “power mediators.”Footnote ⁵⁸ Or, in Alex Stone Sweet’s more abstract formulation, a dyadic relationship between two parties transforms into a triad consisting of two parties and an adjudicator.Footnote ⁵⁹

To fulfill its umpire-likeFootnote ⁶⁰ role in the triad, an adjudicator must be impartial toward the parties and hence cannot be a “part” of either disputant. Typically, that separation between the impartial third party and the disputing parties is conceived of as independence.

The relationship between impartiality and independence is not entirely intuitive. Whereas we might use the two concepts interchangeably in common parlance, they denote different aspects of adjudication. Independence denotes a predominantly organizational and external criterion whereas impartiality refers to a mainly functional and internal one. For example, in its decisions on Polish judicial “reforms,” the European Court of Justice (ECJ)’s judges interpreted impartiality as an adjudicator’s fair-mindedness and objectivity.Footnote ⁶¹ Adjudicators are therefore impartial if their staff make decisions uninfluenced by considerations other than those based on the relevant norms. In this idealized view, (adjudicatory) impartiality is clearly a myth. Implicit norm-derived sentiments are so embedded in our thinking, language, and informational landscape that “truly” impartial decision-making seems impossible.Footnote ⁶² A pragmatic solution, therefore, is to confine elements and influences that might unduly affect decision-making and at least maintain the impression of a decision’s impartiality and safeguard its norm-based verifiability through normative reasoning, which ties adjudication to the existing normative backdrop.

In turn, independence can be seen as the adjudicating organization’s autonomy from outside pressure or factual constraint.Footnote ⁶³ In other words, independence is the organizational and somewhat objectifiable means toward the goal of impartial decision-making.Footnote ⁶⁴ Contemporary discussions mainly revolve around these organizational aspects, summarized as judicial independence.Footnote ⁶⁵ On the most fundamental level, independence demands that adjudicators be free from any external command, hierarchy, or intervention. As the ECJ formulates it:

Although its precise contours vary, independence requires a guarantee that there are independently derived, specified procedures for appointment and nomination, fixed tenure and remuneration of the adjudicating personnel, removal for cause only, a preset budget and, crucially, no encroachment on the power structures an adjudicator is tasked with controlling.Footnote ⁶⁷

Yet from a historical perspective, adjudication’s foundational criterion is impartiality, not independence. Adjudication of private disputes (but to a degree also for disputes between individuals and the organized and powerful, yet private nobility)Footnote ⁶⁸ predates independent judiciaries in the modern sense. Also, as various judicial institutions in the Catholic Church or imperial China indicate, courts as institutions are not limited to constitutional power structures in liberal democracies.Footnote ⁶⁹ That being said, adjudicating individual rights against a given context’s dominant power structure, instead of merely smoothing over social and commercial disputes, remains largely exclusive to the liberal project. Individual rights adjudication, in that sense, reflexively relates to the very emergence of modern public law itself.Footnote ⁷⁰ Even though, as Mauro Cappelletti highlights, “nobody would deny that even at the time of the absolute state … there were nevertheless judicial decisions,”Footnote ⁷¹ absolutist adjudication’s liberty-impairing and erratic nature ended up necessitating and enabling an independent judiciary and individual rights adjudication.Footnote ⁷² Using individual rights as power mediators requires disentangling the accountability mechanism from the encroaching power structure. Metaphorically put, a leash is no use if nobody is holding it.

The degree and design of independence remain contingent on each adjudicator’s organizational setup.Footnote ⁷³ As Cappelletti says, “[T]here exist, of course, thousands of definitions of the terms ‘court,’ ‘judge’ and ‘judicial,’ and no one who has a comparativist’s mind would insist on any of them as absolute.”Footnote ⁷⁴ The focus on an independent judiciary permeated through Europe and the United States after the American and French revolutions and the ensuing adoption of the principle of tripartite government.Footnote ⁷⁵ As a means of controlling public power in general as well as the burgeoning forms of administrative power, judicial independence stabilized external accountability but also created frictions between idealized notions and practice. Debates are legion, from the much discussed “counter-majoritarian difficulty” to courts’ legitimacy and practical ability to control the expanding administrative state’s “independent” agencies and informal acts.Footnote ⁷⁶ The control of administrative power by adjudicating individual rights is never detached from public power or, in the case of administrative adjudicators, not even independent of administrative power structures.Footnote ⁷⁷ In these contexts, adjudication was itself an exercise of public power.

To reconcile impartial and independent adjudication with the practical need to finance, staff, and (democratically) legitimizeFootnote ⁷⁸ adjudicators, public-law adjudicators developed various, graded stages of adjudicatory independence. Courts and the judiciary are not the sole torchbearers of the judicial function. Rather, many systems such as those of the European Union, the United States or England are familiar with sites of “administrative justice” revolving around adjudicators that are not courts.Footnote ⁷⁹ These adjudicators often lack the same independence safeguards as those enjoyed by “classic” courts.Footnote ⁸⁰ But such non-court adjudicators are, as a result of their respective legal and political systems’ peculiarities, endowed with “sufficient” independence safeguards.Footnote ⁸¹ For example, members of EU Boards of Appeal also have fixed tenure, can only be removed for cause, are not entrenched in “their” agency’s hierarchical apparatus, and are required to make decisions independently and impartially.Footnote ⁸² Particularly for adjudicators that provide accountability to administrative power structures, tensions between impartial decision-making, organizational independence, and structural entrenchment are a necessary side effect of the diffusion of judicial functions beyond the traditional judicial system.Footnote ⁸³ Assessing an adjudicator’s independence thus remains necessarily context-sensitive and flexible.

2.2.5 Rules

Once an individual who seeks to protect their rights has turned to an impartial and independent adjudicator, the adjudicator cannot decide at whim but only based on a particular normative framework.Footnote ⁸⁴ This normative framework must meet several requirements. On the one hand, it should be preset. This means that rules should not be made ad hoc but should reflect the relationship between the individual and the power structure as it was when the alleged rights-infringement took place. The preset rules upon which the adjudicator bases its decisions can be of various origins; typically, they will be the law of a given state, public organization, or stipulations contained in international law. Further, it is crucial that the rules are – at least in part – external to the controlled entity. For example, adjudicators control whether administrative agencies comply with external provisions like the EU treaties, the US Constitution, the agency’s founding regulation, or the relevant act of Congress. The adjudicator then applies the facts of the case, in terms of how the power structure exercised its power, to the normative material at hand – that is, the law regulating the agent’s powers – and then decides whether the agent’s acts conform with this normative material or not.

If the normative material violates higher-ranking rules, individual rights adjudication may also include reviewing the normative material itself. For example, if the adjudicator considers statutory provisions to conflict with constitutional requirements, depending on its authority, the adjudicator might declare the lower-level rules void or inapplicable. However, whereas reviewing rule enforcement is the bread and butter of individual rights adjudication, reviewing rulemaking is a more complex endeavor.

2.2.6 Outcomes

When an individual goes to an adjudicator to enforce their rights and control power, adjudication may yield several outcomes. The stereotypical outcome of adjudication – especially if conceptualized mainly as a dispute – is a decision. A decision assesses the normative material and typically orders the losing party to commit or omit a certain act. For example, a court may decide that an agency’s approval for specific medication did not follow the required procedural steps and violated the relevant normative material. The court would then annul said approval and perhaps order the agency to examine the approval again, following all the necessary procedural steps. Thus, the role of individual rights adjudication is typically to review and control decisions made by (administrative) actors – and not to make a new “administrative” decision.

In a historical context, an adjudicator’s decision is typically legally binding for both parties, especially the public agent. In constitutional democracies, public power is conditioned by the rule of law. In fact, the ability to make legally binding decisions constitutes perhaps the crucial feature distinguishing adjudicators who really control public power, like courts, from other more marginal guardians that provide more circumscribed services like mediation, arbitration, expert review, or transparency.Footnote ⁸⁵

However, from a functional perspective, we should not overemphasize the importance of a decision’s legally binding nature, because the efficacy of legally binding decisions is conditioned by the degree to which the controlled actor abides by the rule of law. Reversing this formulation, legally binding decisions are only effective in rule of law systems whereas less stable, novel, and still emerging normative systems require other mechanisms to safeguard a decision’s real-world effectiveness.

Even within stable rule of law systems, adjudicators use a variety of instruments to control organizational behavior. Binding decisions are an important instrument, but by no means the only one. As argued above, disputes about individual rights often merely reflect more deeply rooted governance problems. Such problems could manifest as recurring patterns or institutionalized structures within an organization that inevitably and persistently produce individual rights violations. We can identify various examples whereby adjudicators addressed deeper organizational patterns that had led to various kinds of individual rights violations. Among the most notable are racial bias within police departments,Footnote ⁸⁶ administrative inadequacies leading to “systemic deficiencies” in EU Member States’ asylum systems,Footnote ⁸⁷ or the lack of political will to take measures necessary for meeting carbon reduction targets.Footnote ⁸⁸ In such cases, judges often act in ways that are “forward looking, fashioned ad hoc on flexible and broadly remedial lines, often having important consequences for many persons including absentees.”Footnote ⁸⁹ This is because correcting structurally flawed patterns within a powerful organization, especially administrative ones, is ultimately “not a dispute between private individuals about private rights, but a grievance about the operation of public policy.”Footnote ⁹⁰ Remedying structural patterns or, in Abram Chayes’ words, to “operate public policy,” adjudicators resort to flexible, more forward-looking remedial approaches. These can be more effective than singular, backward-looking, but binding decisions. In all the abovementioned cases – despite their variety – the adjudicator outlined future handrails for the controlled organization that would remedy the underlying governance issue. In COPPAR v Rizzo, an example I borrow from Chayes, a District Court required a police department to “submit … a comprehensive program for dealing adequately with civilian complaints alleging police misconduct.”Footnote ⁹¹ Such a program should, for example, safeguard “[a]ppropriate revision of police manuals and rules of procedure spelling out in some detail, in simple language, the ‘dos and don’ts’ of permissible conduct in dealing with civilians.”Footnote ⁹² In N.S., the ECJ described how a Member State could assess whether “systemic deficiencies” imbue the asylum procedures in another Member State, namely by resorting to “regular and unanimous reports of international non-governmental organizations bearing witness to the practical difficulties in the implementation of the Common European Asylum System.”Footnote ⁹³ The Bundesverfassungsgericht (Germany’s constitutional court) ordered the German legislator (!) to ramp up its carbon reduction plans “according to the guidelines [Maßgaben]” formulated by the Bundesverfassungsgericht.Footnote ⁹⁴

Taking a normative perspective on the separation of functions, adjudicatory review of the deeper flaws of power structures seems intuitively more contentious than superficial remedies of individual rights violations.Footnote ⁹⁵ However, drawing the line between “effective” adjudication that tackles the root of governance problems on the one hand and misguided adjudicatory creep into the realm of policymaking on the other requires fine-tuned measuring instead of principled simplifying.Footnote ⁹⁶ Adjudicatory efficacy depends on an adjudicator’s authority within its legal system, the applicable normative framework, and other control tools at hand.Footnote ⁹⁷ Hence, a clear-cut, generalizable normative solution as to what exactly constitutes effective judicial protection arguably does not – and cannot – exist.

For the purposes of this book, we can conclude that adjudicators use various tools to exercise control over large scale organizations, ranging from binary, binding decisions to open-ended, flexible recommendations, which are of course reinforced by most adjudicators’ ability to issue binding decisions in cases of noncompliance. Therefore, to ensure that adjudicators’ influence over structural reforms at large organizations can be properly wielded, binding decision-making powers serve as a sword of Damocles lending weight to non-binding normative sentences issued by the adjudicator.

2.2.7 Appellate Review

As we are jumping back and forth between Europe and the United States, another terminological note seems warranted. The term “adjudication” has different connotations in England (and most other countries that follow common law) than in the United States.Footnote ⁹⁸ In England,Footnote ⁹⁹ adjudication refers to third party appellate review of administrative decisions.Footnote ¹⁰⁰ In contrast, in US parlance, adjudication broadly refers to independent decision-making related to administrative decisions as well as to their appellate review by a third party.Footnote ¹⁰¹ This broader American concept of adjudication reflects the US’s Administrative Procedure Act of 1946, which applies adjudicating as an umbrella term for various types of administrative decision-making.Footnote ¹⁰²

Here, our concern is with the ex-post control of platforms’ administration-like rulemaking and, especially, rule enforcement. Thus, we use the narrower understanding of adjudication as appellate review as this is more common, especially in England and many of its former colonial areas (such as Canada or Australia).

2.3.1 Constitutional or Administrative?

It is by now a truism that some of these digital corporations – or whatever comes after them – became so pervasive, inescapable, and powerful that we must think about them in public law terms.Footnote ¹⁰⁵ Public law imagery is so pervasive that it transcends political divides. Liberal voices call for expanded rights, adjudication, and procedural safeguards, while increasingly vocal “conservative” critics, especially since the 2024 US elections, demand greater majoritarianism, denounce alleged viewpoint discrimination, and rail against “woke” content moderation policies.

This book focuses on the first wave of public law arguments: rights, adjudication, and due process. Yet many of the critiques leveled against these frameworks are themselves rooted in public law reasoning. In a sense, we are witnessing a reenactment of familiar public law debates – countermajoritarianism, freedom of speech, and institutional design – now transposed onto the realm of private and regulatory governance.

However, in general and scholarly discourse, these public law terms are often framed as constitutional rather than administrative. That is especially the case for discourses about content moderation. In principle, that dominance of constitutionalist discourse about platforms is not surprising. Content moderation intuitively links to freedom of speech, which is often understood as a quintessentially constitutional issue.

However, this book takes a slightly different approach. I emphasize the administrative and bureaucratic elements of much of content moderation and, for the fact of the matter, private governance in general. This complements the descriptive insights and prescriptive ideas by digital constitutionalists.Footnote ¹⁰⁶ The motivation to focus on the administrative instead of the constitutional dimension comes predominantly from an empirical, rather than a solely normative perspective. Of course, many norms of constitutional law like non-discrimination or a minimum of legal coherence as it is implicit in any rule of law-based system would improve platform governance. However, if one looks at how content moderation actually looks like, analogies to administration and bureaucracies appear much more obvious. Platforms design, regulate, and police infrastructures with eminent importance for public life. They enact a myriad of extremely detailed norms – much more detailed than “free expression” – and devise the most intricate specifics of their technological infrastructures. Platforms organize these practices globally, partly accommodating to specific normative needs of local jurisdictions, partly striving for global coherence to maintain their economies of scale. These elements entail choices and practices that could be described as hierarchy and deviation, deferral and review, coherence and flexibility. Seen in the abstract these normative and organizational dilemmas remind us of structural features of the administrative state. In that sense, content moderation – especially through its co-optation by regulators – becomes a globalized transnational yet privately dominated administrative space.

Crucially, I use the notion administration or administrative law broadly. It refers to the acquis of transnational administrative law as we know it from the European Union or global governance structures but also domestic administrative law. In that sense, I particularly draw inspiration from the body of theories devised in the 2000s as “Global Administrative Law” (GAL) initiated by Benedict Kingsbury, Richard Stewart, and Nico Krisch. Importantly, (global) administrative law is now also met with great interest by private lawyers. Especially the work of Rodrigo Vallejo on “Private Administrative Law” highlighted the administrative elements in private standardization mechanisms and administrative law’s potential beyond the ubiquitous but epistemologically laden “governance” paradigm.Footnote ¹⁰⁷ Others have rightly highlighted the relevance of standards, transnational private governance, and private ordering.Footnote ¹⁰⁸

Coming from that empirical starting point, focusing on the administrative and bureaucratic rather than the constitutional dimension of content moderation and private governance offers several advantages. To begin with, administrative law is more detailed than constitutional law. Freedom of speech in an individual case is a constitutional issue about balancing one fundamental right against another. In contrast, governance of millions of speech acts becomes primarily an administrative issue as said balancing exercise hinges upon effective and faithful implementation down the line. Constitutional law’s brushstrokes are often too broad for the phenomenally complex and thickly layered content moderation bureaucracies we have today. Surely, procedures ought to be fair and nondiscriminatory and align with all kinds of constitutionalist values. But what does that mean for the organizational and procedural structuring of contemporary moderation behemoths? Especially so if our political economy protects platforms’ profitability in often very absolute terms. These companies direct tens of thousands of human moderators, lawyers, engineers and, to a still ill-understood extent, automatic decision-making. Administrative law offers answers to these questions.

In addition, administrative law comes with way fewer side effects than constitutional law. Applying constitutional law notions descriptively to private actors like Meta may be justified through the latter’s humungous power. But doing so may remain a bad description as most to all big social media companies are anything but “constitutional” in a liberal sense.Footnote ¹⁰⁹ Neither do they feature meaningful participatory – not to say democratic – elements, nor do social media platforms meaningfully self-restrict their power or value the “public interest” higher than their own interest in profit. To be clear, that is not at all meant pejoratively. It is just the very purpose of a company. However, adorning companies too quickly with the label of being “constitutionalized” may cloak them in a legitimacy they do not deserve. Administrative law on the other hand, although also developed in the context of public power, comes with way less legitimatory grandezza. Administrative law is often perceived as boring. That is, in this case, an asset. Saying that platforms administrate is descriptively accurate, as they deal with millions of customers and do so – intriguingly – increasingly more informed (if not necessarily guided) by a public law understanding of what good administration entails.

Emulating individual rights adjudication for content moderation is expected therefore to achieve a similar goal as historic individual rights adjudication achieved vis-à-vis administrative actors: protecting individual rights, advancing accountability, and controlling otherwise largely unchecked power structures.

2.3.2 Platforms

To substantiate this plea for a more administrative-focused perspective on content moderation, platforms, and private governance in general, I proceed with a quick look at the phenomenon of the platform. Although predominantly economic, it gives rise to social and normative implications. The term “platform” is a metaphor relating to a specific way in which data is presented and economic transactions are conducted in the information economy.Footnote ¹¹⁰

In the abstract, a platform can be conceived of as an intermediary, virtual “structure” that allows people to communicate publicly.Footnote ¹¹¹ Remarkably, the very term “platform” epitomizes the inevitable performative spin of any discourse about or practice of emergent digital power structures. As Gillespie has argued, companies have been strategically advancing the term “platform” to promote ideals of egalitarian connectivity at least since this century’s first decade.Footnote ¹¹² By framing companies like Facebook or Amazon as “platforms” and not, for example, as “online wholesale outlets” or “data mines,” the metaphor implies the very criteria by which those companies will be judged.Footnote ¹¹³

We can distinguish platforms from other communication tools like messaging, email, or chatrooms by their degree of public availability, accessibility, and direction of communication. Although the details vary, the underlying idea of a platform is to facilitate contemporaneous communication among a group of people. Communicative acts on platforms are “public” in that they are multidirectional since communication on platforms is often directed to a specific or unspecified audience. In contrast, emails or chats tend to be bidirectional because communication often occurs along the lines of existing social relations and, crucially, cannot be accessed or detected by other program users.Footnote ¹¹⁴ The term “platform” also entered black letter law, with the EU’s DSA defining the platform as an online infrastructure that “stores and disseminates information to the public” at the request of users.Footnote ¹¹⁵

On some platforms, like Amazon or Alibaba, interaction is mainly commercial and based on trading. This type of platform can be seen as a marketplace bringing together buyers and sellers.Footnote ¹¹⁶ On other platforms, such as X, formerly Twitter, communication is generally social, cultural, or political. Many platforms blend elements, for example Facebook, which is primarily a social network but also includes more transaction-oriented features on its site. Even gaming or pornography sites can be seen as platforms as many offer channels of communication or even monetary exchange.

Platforms that focus mainly on social, cultural, or political communication are called social media or social networks and are sometimes described as a “digital marketplace of ideas” or, in reference to Habermas, a “digital public sphere.”Footnote ¹¹⁷ However, the events of the last decade emphasize the deeply ambivalent effects of social media.Footnote ¹¹⁸ On one side, social media platforms have helped to foment discourses and social movements that have upended or at least shaken normatively questionable power structures.Footnote ¹¹⁹ Yet social media has also enabled and intensified discourses and movements that many view as normatively problematic, such as misinformation and anti-vaccine sentiments during the Covid-19 pandemic, as well as a slurry of xenophobic, racist, misogynous, antisemitic, or otherwise hateful content. To a degree, social media’s “democratization” of discourse has given airtime to unsavory viewpoints that were formerly hard to access by most people because editorially mediated news outlets tended not to feature these kinds of content. It seems almost ironic that many now point to Habermas when lamenting the demise of twentieth century mass media – television, newspapers, radio – as Habermas himself held those very outlets responsible for what he described as the demise of the public sphere when compared to direct, face-to-face debates in bourgeois European circles in nineteenth-century Vienna, Berlin, London, and Paris. But aristocrat-studded salons have their own problems. Whether nineteenth-century discussions in salons and at dinner tables – all male and with a lot of smoking, I suppose – were indeed preferable to what we have today may be ultimately a matter of taste. Either way, idealizing phenomena we did not experience ourselves – especially historic ones – often risks idealizing a narrative that reflects only a fraction of what in and of itself is a rather complicated and ambivalent phenomenon.

Last, we should not misinterpret the notion that social media is about social, cultural, or political communication. Communication on social media platforms is inherently commercial. Many social media companies rely on advertising, which is personalized and then placed on their sites and those of others. Social media companies “mine” the behavioral data of their users and exploit it for advertising, profiling, and other forms of monetization.Footnote ¹²⁰ Against this backdrop, any communication, even someone’s not explicitly communicative behavior or activity, is of commercial value to the company that owns and maintains the platform because it still yields information – data which can then be monetized – about that person.

2.3.3 Design, Regulation, Enforcement

These observations bring us to content moderation, the umbrella term for social media companies’ governance of online communication.Footnote ¹²¹ Content moderation as well as the broader term “platform governance”Footnote ¹²² has been described in detail by others.Footnote ¹²³ For our purposes, it suffices to carve out several key elements of current, and still evolving, conceptions of content moderation.

First, content moderation should be interpreted broadly. It refers to the design and construction of a type of digital infrastructure – the platform – and the devising of rules that govern the operations of that infrastructure as well as the enforcement of those rules. As argued above, content moderation comprises at least three elements: design, rulemaking, and rule enforcement. As Kate Klonick’s influential piece in the Harvard Law Review argued, the governance by “new governors” of online speech “bear[s] remarkable resemblance to heuristics and structures familiar in legal decision-making.”Footnote ¹²⁴ These rules prohibit certain kinds of communication and are unilaterally devised by social media companiesFootnote ¹²⁵ as well as often bleeding into various other contexts. As we will encounter those rules repeatedly in the empirical parts of the book, we will not go into greater detail here.Footnote ¹²⁶

But content moderation is more than just rulemaking and rule enforcement.Footnote ¹²⁷ Content moderation also refers to the design of a platform and its business model.Footnote ¹²⁸ The type of content a platform’s algorithm amplifies, the data on which a platform’s algorithms base their operations, and the many other normative decisions and technological issues involved in creating and operating a platform all, one way or another, moderate the content that users access.Footnote ¹²⁹

Second, platform power in content moderation is largely unilateral and generally unaccountable. Although social media platforms operate within legal systems, their leeway in designing content moderation systems is broad. Sheltered by particular conceptions of liability developed in US and European legal frameworks, platforms are considered by their owners to be mere “intermediaries” that are not liable for the content of information posted by their users. Historically, platform owners’ motivation to moderate is not predominantly for legal reasons, as they tend not to be liable for illegal content anyway, but rather on the basis of their own interest in maintaining a (relatively socially acceptable) discursive environment that keeps users on the platform and accommodates the advertisers and potential buyers of the products those ads promote. To keep communication “civil,” platforms therefore began formulating ever more specific rules and intensified the initially human and then algorithmic enforcement of those rules. Simultaneously, platform owners and employees sometimes made conflicting choices, designing their offerings in such a way as to amplify offensive content, which they discovered would intensify user engagement more than “positive” content.Footnote ¹³⁰

Third, as mentioned above, we should think of content moderation as an exemplar. Clearly, social media is a relatively novel phenomenon. Nonetheless, private rulemaking and private governance of communication, resources, and people is by no means new. Although the future picture of world politics, trade, and digitization is necessarily blurry, it seems plausible that globally active, vertically integrated corporations with unilateral governance capacities over important sectors, in one form or the other, come what may, will become an ever more prevalent trend. Perhaps even more plausible is that digitization will play a huge role in future governance and that the power of a few, oligopolistic private power structures will strongly shape these developments. In other words, social media platforms and content moderation are not the first, nor will they be the last, power structure in potential need new forms of adjudicative guardianship.

Assuming future policymakers, managers, lawyers, and other experts might again come up with the idea to emulate individual rights adjudication to control private power, this study seeks to keep this overarching context in mind. For that reason, I do not speak of content moderation all the time but also about, for example, private governance or private ordering.

2.3.4 Private Content Moderation

After outlining these definitional contours, we now move on to dissect the public and private aspects of social media platforms’ content moderation. We begin with the private aspects.

Social media platforms like Meta or X are private organizations – in other words, they are almost all pivately owned and controlled.Footnote ¹³¹ Generally speaking, one can portray the relationships between these private organizations and hundreds of millions of individuals as solely private. A corporation offers a service and consumers access and use it, paying mainly by making their private data available to the company.Footnote ¹³² In that sense, controlling the private bureaucracies of platforms is only “customer service” – a characterization used by an interviewee working for Meta’s Oversight Board.Footnote ¹³³ Correspondingly, we see many private-law elements in Emulated Guardians’ formal structure. For example, corporate law imbues the Oversight Board’s formal structure, set up by a noncharitable trust based in Wilmington, Delaware.Footnote ¹³⁴ Similarly, one core ideational role model for the DSA’s out-of-court dispute settlement bodies was alternative dispute resolution in consumer relations. As we will see in Chapter 3, the European Commission modeled its proposed adjudicators that were supposed to “tame the giants” in parts on oversight bodies that, for instance, enforce passenger rights in disputes with airlines or resolve contractual disputes between customers and car insurers.Footnote ¹³⁵

However, analyzing the power relation between social media platforms and individuals based only on their formal structure obfuscates the functional dynamics at play.Footnote ¹³⁶ As Deirdre Curtin and Linda Senden argued, “[t]he distinction between private and public regulation is, in any event, blurred and unclear, with traditional public administration becoming ever more ‘unbounded’ as outside actors become involved to varying degrees.”Footnote ¹³⁷ So “private law is public law in the sense that private regulation is part and parcel of democratic self-rule by a collective.”Footnote ¹³⁸ In turn, privately organized actors, like platforms, are subject to public-law scrutiny as they exercise power as it affects people on issues related to public law, such as freedom of expression. Yet, in our context, an idealized picture of private rulemaking as “democratic self-rule by a collective” is euphemistic. The rules governing the relationship between social media platforms and users are formal contracts. But a meeting of minds when it comes to such contracts is pure fiction; their provisions are periodically amended and unilaterally dominated by social media companies. Some researchers have even begun questioning whether social media platforms’ terms of service can even be interpreted as contracts at all.Footnote ¹³⁹ Others highlighted that “the law” stabilizes platform power through accommodating the political economy in which informational capitalism operates. In other words, platforms do not need to break the law as the law itself became part of the power relationship between platform, public authority, and user.Footnote ¹⁴⁰ It would then follow that neither rulemaking by platforms, nor public regulation thereof can be reasonably construed as pure democratic self-government.

2.3.5 Communication and Fundamental Rights

However, despite the private law elements just mentioned, understanding (social media) platforms through a public law lens is now widely accepted. Such a public-law perspective becomes especially relevant whenever a private protagonist acts on the public. Social media platforms touch upon two aspects of crucial public concern. These are, on the one hand, communication and, on the other, fundamental rights, especially freedom of expression.

The space social media platforms design for their users serves as a forum for public and private communication. The importance of communication as a formative source of the self and as a cornerstone of society and politics is well-known.Footnote ¹⁴¹ Society and politics are products of communication, which perhaps points to the core of the human condition.Footnote ¹⁴² People are “political animals” by virtue of their ability to communicate.Footnote ¹⁴³ Hence, enabling and regulating communication – as social media platforms like Meta and X do – means enabling and regulating a key ingredient of the social and political order we inhabit. Later theorists of justice,Footnote ¹⁴⁴ deliberative democracy,Footnote ¹⁴⁵ or the linguistic “turn”Footnote ¹⁴⁶ have emphasized the importance (and fragility) of communication. Therefore, the governance of people’s communication on social media platforms will inevitably warrant public-law scrutiny. As Linnet Taylor put it,

Second, communication’s value is not only a philosophical or sociological truism but has long been established in law. The fundamental right of freedom of expression protects online communication.Footnote ¹⁴⁸ Further, many specific fundamental rights, ranging from freedom of religion, and association, to the freedom to conduct business as enshrined in the EU’s Charter of Fundamental Rights, make up a panoply of applicable rights. Platforms have become indispensable forums for the exercising of such rights, especially freedom of expression but also freedom of religion and the freedom to conduct business.Footnote ¹⁴⁹ As the EU’s DSA states in its preamble, to “exercise their fundamental rights,“ EU citizens are dependent on “responsible and diligent behavior” by platforms.Footnote ¹⁵⁰ Public and especially administrative law offers a toolkit for normatively evaluating the organizational behavior that enables and sometimes restricts fundamental rights. Albeit platforms are not directly bound by fundamental rights, their actions shape how others exercise theirs. Further, platforms’ gatekeeping function may impose procedural duties on them, such as those outlined by Germany’s constitutional court in its Stadionverbot decision.Footnote ¹⁵¹ In the United States, in contrast, the Supreme Court appears not at all poised to envelope social media platforms in a denser net of obligations of how to moderate the content of their users.Footnote ¹⁵²

2.3.6 Regulatory Administrification

However, as important as communication and fundamental rights are, they cover only the subject of platform power but not its functionality. Because platforms deal with “public goods” like public communication and fundamental rights at a massive scale, they are increasingly regulated to abide to public law norms. These norms regulate how platforms shall deal with their users. The normative model for many such rules – fundamental rights, due process, proportionality, nondiscrimination, hearing rights, duties to give reasons, and so on – is public law, or more precisely administrative law. I call this the administrification of platform governance.

There are two key drivers behind platforms’ increasing administrification.Footnote ¹⁵³ These are, on the one hand, regulatory co-opting of platforms and, on the other, increasing public scrutiny and demands for heightened platform accountability.Footnote ¹⁵⁴ Social media platforms represent what the GAL project has characterized as hybrid intergovernmental-private administrations (in the case of the endowment to a public mandate) or entirely private administrators.Footnote ¹⁵⁵ Such private governance organizations, like companies, “have adopted certain procedures of accountability and review in order to enhance their effectiveness and legitimacy; these may have parallels in domestic administrative and private law that are so far underexplored.”Footnote ¹⁵⁶ Hence, studying platforms as “part of global administration [allows us] to trace similarities as well as differences in mechanisms of accountability developed for public and private bodies.”Footnote ¹⁵⁷ This section outlines how regulatory co-option, which will only intensify with the newly enshrined DSA, incentivizes platforms to adopt administration-like procedures, rules, and formal structures. So far, except for remarks by Balkin, Bloch-Wehba, and Douek, this perspective was not explored in the literature.Footnote ¹⁵⁸

In 2016, several social media platforms, including Facebook and Twitter, now X, agreed with the European Commission on a code of conduct when it came to countering illegal hate speech online.Footnote ¹⁵⁹ The social media platforms agreed to “have in place clear and effective processes to review notifications regarding illegal hate speech” and committed “to review such requests against their rules and community guidelines.”Footnote ¹⁶⁰ Since platforms operate at scale, their commitments toward EU authorities are likely to affect, and perhaps water down, their global standards. In this context, Anu Bradford argued in her eminent book that platforms’ normative material (the so-called community standards) increasingly mirrors the normative acquis of EU law.Footnote ¹⁶¹ However, it remains unclear whether their terms of service really “reflect the European standard of hate speech” as Bradford argues, or merely echo the overall tone of EU law rather than reflecting the EU acquis in any meaningful way.Footnote ¹⁶² Bradford points out that we do not know how these commitments are put into practice; especially since platform rules consist of many sublayers and platform lawyers “ultimately use their own judgment on what constitutes illegal hate speech. The outcome is therefore less likely to be perfectly aligned with the Commission’s regulatory approach.”Footnote ¹⁶³ Building on the preexisting division du travail between social media platforms and European regulators, the Commission proposed a major overhaul of parts of the governance of social media platforms in December 2020.Footnote ¹⁶⁴

2.3.7 Especially, the Digital Services Act

This overhaul, the DSA, reiterated that platforms would be relatively free to devise their own procedures, decision-making processes, and substantive rules if they orient their dealings toward criteria like “objectivity,” “proportionality,” and “due regard for fundamental rights.”Footnote ¹⁶⁵ The DSA imposed several administrative law-inspired duties on platforms, for example the duty to give reasons for decisions against their users (Article 17 DSA) and to “act in a diligent, objective and proportionate manner in applying and enforcing” their terms of service (Article 14(2)).Footnote ¹⁶⁶ The DSA stipulates that platforms should act with

In other words, the DSA infuses public-law ideas and administrative-law procedures into the otherwise private governance of online platforms. In response to this intensifying regulatory framework, platforms ramped up procedures, formulated norms, trained people, and formalized structures that could be characterized as “private bureaucracies.”Footnote ¹⁶⁷ Social media platforms must deal with notices, state reasons, and take down content that might be illegal.Footnote ¹⁶⁸ In some cases, platforms are in effect auxiliary agents of public power.Footnote ¹⁶⁹ In consequence, legal scholarship is increasingly conceptualizing platforms, more specifically content moderation, as systems of administration or even as bureaucracies.Footnote ¹⁷⁰

As a result, Balkin concludes, we observe a “burgeoning dialectic of governing power and public-private cooperation. Private infrastructure companies develop greater governing capacities than ever.”Footnote ¹⁷¹ Indeed, governmental pressure on platforms has grown in recent years. In sum, regulatory frameworks like the DSA oblige platforms to enforce regulatory interests relating to copyright, personal rights or criminal law, for instance, which in practice compels platforms to be functional administrators. In turn, analyzing platforms through the lens of administrative law appears imperative.

2.3.8 Autochthonous Administrification

The previous sections showed how regulatory co-option molds social media platforms into ever more administration-like entities. However, regulatory endowment is not the only driver behind platforms’ burgeoning private bureaucracies. Platforms’ emulation of administration-inspired rulemaking and rule enforcement reflects demands from users – and, perhaps more importantly, ad-buyers – for “clean” platforms.Footnote ¹⁷² Users and investors demand a platform that is not littered with hate speech, sexism, or xenophobia.

To achieve this, platforms established their own rulemaking and enforcement regimes to monitor and regulate the communications of their users.Footnote ¹⁷³ Faced with growing demands to take down harmful or violent speech, platform companies morphed from mere technology companies to complex regulatory bureaucracies.Footnote ¹⁷⁴ As Kate Klonick’s extensive empirical work shows, “platforms have developed a system that has marked similarities to legal or governance systems. This includes the creation of a detailed list of rules, trained human decision-making to apply those rules, and reliance on a system of external influence to update and amend those rules.”Footnote ¹⁷⁵ In turn, the intricate work of nudging such complex bureaucracies toward normatively desirable behavior is familiar – and has been well documented for social media platforms, particularly in the area of privacy.Footnote ¹⁷⁶

This private rulemaking appears deficient from an administrative law perspective or, in fact, any normative position linked to a notion of self-governance and publicness. Although platforms’ private rulemaking is comprehensive, it remains incoherent and imprecise.Footnote ¹⁷⁷ Still, social media platforms devise their rules unilaterally. Rulemaking remains largely decoupled from the “meaningful consent” (to quote Suzor) by the users whose behavior the platforms govern.Footnote ¹⁷⁸ Rulemaking on platforms has been described as autocraticFootnote ¹⁷⁹ and authoritarianFootnote ¹⁸⁰ but, due to its pervasiveness and normative properties, also as “quasi-legislative” or simply as “law.”Footnote ¹⁸¹ Although I am skeptical when it comes to equating platform rulemaking with blurry, legitimacy-seeking notions like “quasi-legislative,”Footnote ¹⁸² the fact remains that platforms have huge influence in the way they frame and present the norms their private owners deem appropriate.

Further, the administration/administrative law analogy does not stop at cases where platforms are co-opted by governments or when they make and enforce “rules.” The analogy also pertains to the way digital infrastructures are designed by platforms. This “administrification” goes beyond formal rulemaking but includes, especially, infrastructure design and programming.Footnote ¹⁸³ What do users see, and when? Do algorithms nudge people to attack and insult other users?Footnote ¹⁸⁴ A good way to think about all these questions is inspecting them through the method and mindset ofadministrative law. Imagine, for example, a big highway. When the local transport authority plans such a highway, it must discuss plans with the public, make an environmental impact assessment, and may not build it right through a national park. Similar design principles are imaginable for social media platforms too.

Arguing for a more nuanced understanding of the many ways platforms influence user behavior and regulate speech, Evelyn Douek has highlighted that it would be wrong to understand content moderation as “just the aggregation of many (many!) binary decisions to take down or leave up individual pieces of content.”Footnote ¹⁸⁵ Instead, Douek argues, content moderation “is a vast system of administration that includes a far broader range of decisions and decision-makers than the standard picture admits.”Footnote ¹⁸⁶ This book further explores the descriptive and normative implications of content moderations (and, by extension, private governance’s) resemblance to administration. So far, the literature only mentioned the resemblance or used administration as a metaphor but did not really dig into the normative “downstream consequences.”Footnote ¹⁸⁷

2.4.1 Defining Emulation

To emulate is to “imitate” and to “strive to equal or to excel.”Footnote ¹⁹⁰ Emulation comes from the Latin noun aemulus, or “rival,” and the related verb aemulare, “to rival” or “to strive after another earnestly.”Footnote ¹⁹¹ In this book we will encounter both meanings of emulation, connoting imitation, mimicry, and copying on the one hand, and innovation, rivalry, and competition on the other. In the social sciences, emulation is a concept used mainly in organizational sociology,Footnote ¹⁹² political science,Footnote ¹⁹³ and, increasingly, law.Footnote ¹⁹⁴ Political scientists and organizational sociologists have studied such patterns in intermingled but slightly diverging contexts, which are combined here to situate emulation in both the political and the sociological contexts of Emulated Guardians.

2.4.2 Types of Emulation

Based on Simmons et al., we can distinguish three types of emulation.Footnote ¹⁹⁵ These types differ in how actors choose what policy or formal structure they deem “appropriate” to emulate. All types are featured in this book. In practice, the types are interrelated and often converge. We distinguish between the “follow the leader” model, expert theorization, and normalization.

The first type is the “follow the leader” approach, in which social acceptance and notions of the “appropriateness” of policy choices emanate from exemplars of leading actors.Footnote ¹⁹⁶ Emulating leading actors’ policies and institutions is common practice in legal policies, especially in comparative law and the EU’s legislative procedures.Footnote ¹⁹⁷

The second type is the emulation of policy choices or institutional designs because expert groups theorize and provide actors with a rationale to implement them.Footnote ¹⁹⁸ Such “expert theorization” develops in epistemic communities that theorize, define, and sometimes lobby for new policy solutions or institutional designs for existing governance problems.Footnote ¹⁹⁹ Here, no particular real-world example has to exist, although what it is that experts theorize about tends to be informed by existing policies and governance practices.Footnote ²⁰⁰

Expert theorization was a strong driver behind the DSA’s out-of-court dispute settlement bodies and the Oversight Board.Footnote ²⁰¹ Decisive design choices for both the EU DSA’s out-of-court dispute settlement bodies as well as Meta’s Oversight Board were lobbied for and developed by scholars. The roles of experts, professionals, and scholars in legal and organizational practice have been extensively studied in organizational sociology and socio-legal scholarship.Footnote ²⁰² Experts’ stances and professional backgrounds will tend to inform their advice to policymakers and executives of private companies. Or, put more bluntly, as one interviewee who advised Facebook on creating the Oversight Board put it:

This type of expert-driven emulation builds on input from experts conveyed to lawmakers or private-sector executives tasked with devising solutions to a particular governance problem. Whereas expert theorization’s “appropriateness” may be more contested than the “follow the leader” type of emulation (two lawyers, three opinions, as the saying goes), expert-advised emulation banks on the social status of experts and their expertise. Another factor at play is the sheer constraint of time and resources faced by legislators. In particular, ambitious and time-sensitive projects like the DSA often rely on the expertise of people who deal with specific issues far more comprehensively than the average legislator can within the limits of their day-to-day.

Third, the emulation of policy choices or organizational designs often happens when the relevant policy choice and organizational design is considered by experts in the field or by the general public as a standard, a necessary operating procedure, or a norm.Footnote ²⁰⁴ The generating of norms, or “normalization,” often builds on previous exemplars, which are then emulated based on the follow the leader model by other actors and on previous or subsequent theorization by experts.Footnote ²⁰⁵ As Simmons et al. note: “Once a new policy reaches a certain threshold, policymakers take it for granted as necessary.”Footnote ²⁰⁶ This is very visible here. Platforms accrue more and more power. Therefore, accountability appears “appropriate.” The “appropriate” form of accountability then, again, appears to be individual rights adjudication – after all freedom of speech is the one constant tune in this overture.

2.4.3 Policy Diffusion

The intellectual history of emulation builds on two strands of scholarship. On one side it builds on research into policy diffusion, which examines how certain policies diffuse around the globe or within federal structures. Eleanor Westney first introduced the concept of emulation to comparative politics in a 1987 study of Japan’s construction of “Western”-style institutions during the Meiji Restoration in the nineteenth century.Footnote ²⁰⁷ To evade colonization, Japan established a Western-style administrative system, legal frameworks, newspapers, and educational institutions. Within several years, formerly insular Japan had overhauled many of its institutional (as well as organizational) structures. Beth Simmons et al. contrast emulation with coercion (involuntary policy choices), competition (policy choices according to market expectations, that is, a general thrust toward liberalization), and learning (policy choices made after lengthy soul-searching, adapted to specific purposes, often inspired by others).Footnote ²⁰⁸ Political scientists describe such processes as policy diffusion.Footnote ²⁰⁹ Empirical research suggests that emulation is among the most common mechanisms of policy diffusion.Footnote ²¹⁰ Actors reproduce preexisting policy choices because they deem them “appropriate.”Footnote ²¹¹ What an actor considers “appropriate” is influenced by what the actor believes internal adversaries, relevant stakeholders, and the general public consider legitimate. As Eleanor Westney noted, “a foreign organizational model may be used [for emulation] to legitimate changes desired by innovators but which require justification by the appeal to the outside model.”Footnote ²¹² Translated to our context: if a policy maker, Meta executive, or external consultant wants to change how the company is run, they might reference convincing outside examples how things are done in other contexts, for example, through “something like a Supreme Court.” Correspondingly, a key denominator for what is considered appropriate and legitimate by actors throughout the case studies are the values and dogmas of two sociopolitical trends in Western governance since the 1950s: juridification and judicialization. Juridification is “a process in which human conflicts are torn through formalization out of their living context and distorted by being subject to legal processes.”Footnote ²¹³ In simple terms, juridification implies that we now interpret more issues as legal than we used to. The ever-increasing regulatory acquis on the domestic, supranational, and international levels provides ample anecdotal evidence.Footnote ²¹⁴ Judicialization, in turn, refers to the institutions that benefit most from juridification: courtsFootnote ²¹⁵ and other adjudicators.Footnote ²¹⁶ The more we see power imbalances as legal issues and the more we “juridify” and break down a governance actor’s legitimacy into atomized rights disputes, the more powerful courts become. Evidence for judicialization has been observed for various systems, polities, and policies, from the United States to the European Union and its Member States, as well as in the workings of international law.Footnote ²¹⁷ Examples range from the establishing of adjudicatory control over the burgeoning administrative state from the late-nineteenth century to landmark cases like Brown v. Board of Education in the United States to van Gend & Loos in Europe.Footnote ²¹⁸ Hence, paired with the sociopolitical and cultural significance of courts in Western discourses and cultures, the appropriate policy to emulate, almost self-evidently, has been adjudicative control over power. Norms and traditions that help to construct ideals of appropriateness may well conflict in detail, often leading to hybrid or adapted emulations trying to balance competing objectives and rationales. Yet policy diffusion is not the only perspective – especially because it does not account in full for why and when emulations fall short or remain uncompleted. For that, in the next part we need to turn to organizational sociology.

2.4.4 Organization Theory

The sociological angle then highlights the role of public legitimacy and the lingering danger of mere sugarcoating or, as Meyer and Rowan put it, ceremonial reforms. Organizations strive to appear more legitimate by conforming often uniformly to public demands of how a good organization should look like. Organizational theorists describe this emulation of what is seen elsewhere and considered “a best practice” – that is, this trend toward conformity – as institutional isomorphism.Footnote ²¹⁹ That strand of work, soon called “neo-institutionalist,” emerged in the 1980s as a response to then-dominant “rational choice” theories in sociology (and economics).Footnote ²²⁰ Rational choice perspectives posited that “efficient” organizational patterns prevail, whereas inefficient patterns fade away. In contrast, neo-institutionalists argue that many organizational patterns and structures are not necessarily “rational” in terms of efficacy but reflect social and culturally constructed patterns, models, ideas, and a consensus regarding what constitutes a “good” and “legitimate” organization.Footnote ²²¹ Since organizations compete for “political power and institutional legitimacy,” they seek to appear as legitimate as possible.Footnote ²²² One common and relatively easy way to do so is to emulate institutions and organizations that are widely accepted as legitimate. This is why corporate actors have established a panoply of guardianship-oriented institutions ranging from compliance departments to auditors, diversity reports or, as this book shows, self-styled adjudicators.

Notably, emulation often leads to ceremonial changes in the formal structure instead of deep-rooted changes in “policy or strategy.”Footnote ²²³ Emulating ceremonies and formal structures is more easily observable by the public and is perhaps for this reason more effective from a self-regulatory perspective, even though more complex policy changes may be “less easily noticed.”Footnote ²²⁴ Again, the reason is that a ceremonial add-on like a new adjudicator promises immediate societal legitimacy. In contrast, a lengthy overhaul of internal governance practices may go unnoticed by the public.Footnote ²²⁵ Hence, as Meyer and Rowan argued: “the formal structures of many organizations … dramatically reflect the myths of their institutional environments instead of the demands of their work activities.”Footnote ²²⁶ In simple terms, the myth this book investigates is that ever more adjudicators bring more justice.

The “myths” of out-of-court dispute settlement bodies and the Oversight Board’s institutional environment arise from the sociocultural significance of individual rights, judicial review, and widespread belief in an ever-increasing “judicialization” of politics and governance.Footnote ²²⁷ Empirical data compiled by sociologists for decades suggests that organizations use emulation to establish ceremonial, formal structures that are decoupled from the organizations’ work requirements. Doing new things, building fresh formal structures, and (metaphorically) putting on a new ceremony may merely serve to project a legitimizing image onto essentially unchanged practices. From the start, organizational sociologists highlighted the risk of decoupling between emulated, ceremonial structures shown to the outside and organizations’ internal practices which often did not conform to the narratives characterizing the organization’s public image.Footnote ²²⁸ Meyer and Rowan argued: “To maintain ceremonial conformity, organizations that reflect institutional rules tend to buffer their formal structures from the uncertainties of technical activities by becoming loosely coupled, building gaps between their formal structures and actual work activities.”Footnote ²²⁹ Such decoupling serves as a clear warning. If out-of-court dispute settlement bodies or the Oversight Board were merely ceremonial, formal structures decoupled from the activities they were supposed to control, then they could be considered a failure from a public law perspective.

2.4.5 Legal Transplants

Lastly, we should distinguish emulation, as presented in political science and sociological scholarship, from legal transplants, a notion from the field of comparative law. Legal transplants are normative regimes originating in a particular legal system that is then adopted by other legal systems.Footnote ²³⁰ These political science and sociological viewpoints have two advantages over a legal transplant perspective. First, legal transplants refer chiefly to legal norms and not to organizations. Even though more recent scholars studying legal transplants have taken a more context-oriented perspective, the key question is still how a specific set of norms fit into another legal system.Footnote ²³¹ Here, however, we mainly see evidence of the emulation of a formal structure – for example, the Oversight Board and Article 21 DSA emulating adjudicators – but not necessarily the “transplanting” of legal norms. On the contrary. As we will see in the case studies, many legal norms that might be of great value are not “transplanted” to their new context in private governance. Perhaps the most crucial aspect is the adjudicators’ authority. In any serious rule of law system, it is “axiomatic” (using Paul Craig’s word) that adjudicators protect individual rights, in one way or another, by bindingly deciding disputes between individual and public power structures.Footnote ²³² However, the authority of both the Emulated Guardians studied here to bindingly decide cases concerning the activities of social media platforms is severely limited. Hence, perhaps the most fundamental rule of individual rights adjudication has not been successfully transplanted into private governance. If this is the case, then out-of-court dispute settlement bodies and the Oversight Board are the opposite of legal transplants. They transplant an institution – the adjudicator – without transferring the underlying norms. Chapter 5 details how this incomplete emulation impedes both bodies’ roles and activities as guardians.

2.4.6 “Designing This Thing”

Applied to the Emulated Guardians studied here, we may theorize that their creators – the EU and Meta – emulate existing practices, here: individual rights adjudication, to conform to societal expectations. In other words, organizations emulate other organizations, policies, or institutions because they wish to be perceived as legitimate and because existing policies, role models, and institutions demonstrably work. Bound by time and resource constraints, large organizations – private and public – cannot search for a comprehensive solution for each governance problem. Instead, they “borrow heavily from their neighbors and ideological peers.”Footnote ²³³ As one interviewee said:

Hence, when the EU regulates the internal digital market, or Meta presumably tries to escape regulation, this involves making a policy choice that would be widely understood as legitimate. In the very first memos sketching the outline of the Oversight Board, the Board’s court-like structure was praised as its most important asset because this would make it, literally, “widely understood as legitimate.”Footnote ²³⁵ In turn, the EU equally strove to infuse legitimate procedures into social media companies’ private governance and thus initially sought to subject private power to “simplified legal procedures” that would “fit the nature of content moderation disputes.”Footnote ²³⁶

Creating adjudicatory oversight to protect the rights of individuals is something we might generally consider legitimate. After all, who would object to more individual rights protection? Politicians, managers, and consultants, for all kinds of justified practical reasons, cannot devise an “original” solution for every governance problem. Arguably, few genuinely “new” solutions exist anyway since all ideas build on previous notions and experiences. It is thus natural that decision-makers draw inspiration from the wealth of existing models; drawing inspiration and even reproducing and emulating is a common form of human and organizational learning. Therefore, in principle, emulation is neither “good” nor “bad.”

However, adopting models that have been devised elsewhere requires careful fitting.Footnote ²³⁷ Even the most attractive-looking ready-to-wear garment will not fit if you pick the wrong size. In the next chapters we will assess whether ODSs and the Oversight Board indeed “fit” their new contexts.

2.4.7 The Wrong Role Model?

Further, from a more critical perspective and beyond questions surrounding the power relationships between individual and platform, the idea that more adjudication leads to more justice rests on shaky ground. If emulating judicial control over administrative power is to work, as argued above, it requires an actor in an agreed leadership position to follow or adapt theorization of a functioning model. The specific exemplars and theorizations contributing to Emulated Guardians share their epistemological basis in the rise of courts and rights-based approaches that have arisen since the 1950s in the United States and Europe.Footnote ²³⁸ However, these meta-narratives of “rights-based approaches” enabling individuals as controllers of powerful institutions have been criticized from historical, political, and sociological perspectives.Footnote ²³⁹ Famously, Gerald Rosenberg argued that courts are a “hollow hope,” which often cannot bring about social change. Courts that tend to align, so the critique goes, with public opinion more often than not (as Friedman argues) often fail to implement structural changes. Even landmark cases like Brown v. Board of Education may be pyrrhic victories placating lawyers, activists, and policymakers with the self-aggrandizing idea of changing institutional practice with one normative verdict issued after judicial processes, or ceasing to control public power altogether because of institutional self-interest.

Therefore, the leading exemplars from the state context and their theories championing the power of courts and the normative traction of individual rights are perhaps compromised. In fact, over the short period of writing this book, the approval ratings of the US Supreme Court dipped from a ten-year high of 58 percent in July 2020 to a record low of only 39 percent approval in July 2025. And, to add more anecdotal evidence, whenever presenting the research on this book to a US American audience at least one person would question the wisdom of emulating what more and more people consider a flawed institution.Footnote ²⁴⁰ This does not mean that individual rights adjudication is not a hallmark of legitimate government and deserving of praise and repurposing in new governance contexts. Yet when emulating individual rights adjudication, such criticism provides valuable insights into the parameters, circumstances, and conditions essential for effective adjudicatory control of power.

This is not the place to discuss whether the above critiques are entirely accurate. Yet, the abovementioned criticism provokes the question of whether emulating individual rights adjudication to control platform power is the correct approach at all. Suppose traditionally constituted adjudicators are criticized for sometimes failing to protect individuals in the face of power imbalances vis-à-vis powerful administrative institutions. How can emulated adjudicators, often with markedly inferior authority and legitimacy, fare any better? Are there perhaps other, more innovative forms of controlling power equally able to protect individuals from the largest administrators of the twenty-first century? Evidence from the case studies presented here, especially those in the private realm, therefore require us to ask the question Martin Laughlin bitingly posed for public law in general: does it make sense to base “a political order on individual rights, and therefore paradoxically basing the legitimacy of society on a thoroughly asocial principle?”Footnote ²⁴¹

Chapter 6 engages deeper with these criticisms. But as the first wave of organizational and normative emulations in the form of the DSA and the Oversight Board consolidates itself, we can fairly say that thinking about rights permeates legislation, self-regulation, and academic approaches like “digital constitutionalism.”Footnote ²⁴² Ultimately, the fact that discourses around controlling platform power have homed in on relationships between individualized (and therefore relatively powerless) users and independent, autonomous, dominant platforms shows how individual rights can become merely an affirmatory feature in the operation of power instead of power’s counterbalance.Footnote ²⁴³ In much the same vein, Douek recently called for a more systemic approach to content moderation, arguing that “the current regulatory focus on procedure in individual cases is a mistake.”Footnote ²⁴⁴ Therefore, emulating individual rights adjudication must not appear as a “natural” evolutionary step for the judicial process tailing the locus of power from the public domestic level to the supranational level and then to private actors. Instead, it may equally well reproduce those power dynamics and mechanisms that tend to stabilize existing power structures.