1.1 Setting the Scene
How to control private power beyond the state? How to guard individual rights in an ever more globalized and digitized world, where traditional markers of authority, like state-centered sovereignty, territory, and black-letter law seem to gradually fade? How to discipline bourgeoning private power structures that seem to be dominating ever more aspects of our lives? What role for public authority, particularly the European Union (EU) – a vanguard of transnationalism since the 1960s – in that transformation? Does its regulatory power indeed reverberate into the transnational, globalized space of the informational economy? What exactly are the Union’s normative mechanisms to assert European rights and values, apart from contextual factors like leveraging its formidable market size? Or, in general, what tools remain for state actors to enforce their policies vis-à-vis vastly powerful digital corporations?
So far, a common approach to these issues has been to reuse tried and tested vocabularies, mechanisms, and ideas known for controlling public power, especially those from public law. Those public law ideas are then reused for private governance. Rights, adjudication, due process, majoritarianism, and accountability to name a few. That reusing regularly necessitates creating new versions of these older role models, especially existing institutions. The institutions already in place, for example courts, are often considered to be ill-equipped to counter the scale and velocity of ascendent private power structures, especially globally active, vertically integrated digital corporations like Meta, Google, OpenAI, ByteDance, or Tencent. Hence, academics, politicians, and activists often demand new accountability mechanisms to answer these new challenges. This demand for “new” institutions, especially in private and regulatory governance, however, regularly turns to old ideas for inspiration and, quite inevitably, remains discursively embedded in the dominant cultural, legal, and epistemological paradigms of the time. Therefore, when the topic is controlling large power structures, many such old ideas and discursive frameworks stem from public law. By public law, I mean mostly constitutional, civil rights, and, especially, administrative law. To confront the massively concentrated, infrastructural, and globe-spanning power of some private actors - social media companies, in our case - many turn to the emulation of classic public law mechanisms to keep profit-driven power in check.
This book examines what I take to be a fundamental transformation in the way the normativity of public law ideals – individual rights, formalized accountability, and self-governance – unfolds in an era in which the state, long the central locus of organized authority, is gradually receding. Rather than fixating on beginning or endpoints, the analysis centers on the transition itself – foregrounding the process of emergence. It conceptualizes and critically assesses Governance by Emulation as a key mode of ordering beyond the state, particularly in private and regulatory domains.
Emulation captures the dynamic by which organizational actors – such as the European Union or Meta – construct new policies and institutional frameworks by imitating pre-existing, institutionalized models from other contexts. In this case, those models stem from the familiar repertoire of administrative, constitutional, and civil rights law. In short: when searching for credible governance, many now look to public law for inspiration. However, the nascent institutions emerging from this mode of governance carry a complex legacy. Untangling this legacy provides the book’s red thread
In the following chapters, we will study two particularly prominent examples of this common approach. First, the out-of-court dispute settlement bodies, or short ODSs, as foreseen in the EU’s Digital Services Act as a mix of classic alternative dispute settlement and administrative-style civil rights protector. Second, the social media company Meta’s Oversight Board, hailed as “something like a Supreme Court for Facebook” by Mark Zuckerberg.Footnote 1 Albeit vastly different in detail, both exemplify the same idea: emulating the adjudication of individual rights vis-à-vis a large and often quite bureaucratic power structure. In our example here, that power structure are social media companies, more precisely their content moderation machinery. To put it simply, this book studies how the EU and Meta, quite remarkably, both came up with a novel kind of adjudicator to control content moderation and protect the rights of billions of users – or at least appear to do so.
Those novel adjudicators epitomize the thinking of a whole generation of lawyers. The choice to introduce rights-based adjudication into private and regulatory governance of digital power structures reflects liberal convictions and presumptions about the role of rights in democratic discourse and the power of adjudicators to advance social justice. On a more abstract level, the underlying pattern – novel power structure, public demand for accountability, emulation of public law role model like courts, rights, and so forth – seems so self-evident, so obvious, that we will most likely see it again in future reforms in other sectors.
Building on that, we shall call the concrete form of emulation we study here, that of individualized accountability structures, Emulated Guardians. Guardians in reference to their – purported – accountability function and one of public law’s aporiae: Who guards the guardians? The term emulation then highlights their distinctive emulative nature. These guardians seek to reproduce but also to adjust and, potentially, appropriate the symbolism of preexisting role models. I use the term emulation without connotation. In principle, emulating is neither good nor bad. We will see in the case studies, once these Emulated Guardians begin to practice, they develop a life of their own, aspects of which are positive, others more negative.
Emulative arguments and ideas are pervasive in debates about regulation and self-regulation of corporate power structures. We perceive social media platforms as increasingly dangerous and unaccountable threats to individual freedom, social order, and democracy itself. Again, social media platforms are a placeholder here. We could engage in strikingly similar discussions on topics such as artificial intelligence, privacy, data governance, and (digital) infrastructure, alongside other unfolding economic developments. Looking beyond these already tangible realities, consider the largely unregulated domains of extraterrestrial resource extraction, commerce, and – should it ever come to fruition – space colonization.
Simultaneously, we consider public law and its institutions to be perhaps the best way to contain powerful organizations that might threaten liberty, society, and potentially even democratic governance. Why? Because, in the past, public law delivered. We, or rather the ancestors of some of us, overcame absolutism, fascism, and various other ghastly modes of governance. Public law did play an important part in this. Courts then assumed a prominent role in that process, especially since the late 1950s. Therefore, why not paving new avenues for individuals to litigate their rights vis-à-vis private power structures that are widely likened to authoritarian or feudal regimes?
Because it is not so easy. This book identifies and problematizes the deeply ambivalent consequences of emulation as a technique of ordering, be it through norms, organizations, or else. The pathways explored here are based on ideas from a range of disciplines, insights from textual analysis, conclusions from empirical findings this present work has produced, among these – perhaps most illuminating – evidence from expert interviews with people working inside and outside the EU and the Oversight Board, as well as activists, and prior Meta employees. As a result, this book offers fresh perspectives on these practical, normative, and institutional aspects of the discourse and (self-)regulatory practice regarding content moderation and, by extension, private governance in general. Crucially, the book empirically traces how emulative techniques and arguments structure decision-making and institutional design in public regulators as well as private companies. And, the book is, to the best of my knowledge, the first monograph on the Oversight Board and one of the very first on the Digital Services Act. By comparing the Digital Services Act’s ODSs and the Oversight Board, one originating from the EU’s regulatory agenda, the other from Meta’s self-governance, the book shows the varied shades of normativity that canonical public law ideas may take in transnational constellations. The lessons we may draw from this are relevant way beyond content moderation. Holding globalized private power structures accountable is a truly global concern. Today, one of that concern’s most pressing manifestations is content moderation. More will follow.
Emulating public law ideas – and more broadly, using public law as an ideational framework to analyze private power structures in the twenty-first century, particularly digital corporations – transcends political divisions. This book primarily examines the emulation of rights, adjudication, and administrative justice. These are, at their core, “liberal” concepts (in the political theory meaning of the term), especially in an American context. They take center stage here because they dominated the debate on social media governance throughout the 2010s and early 2020s. Yet these ideas have never gone unchallenged. Over time, conservative and more libertarian voices have mounted multifaceted critiques, targeting the speech governance exercised by online platforms in the USA, the regulatory ambitions of the EU, and, more fundamentally, the very choice of values, principles, and institutions operationalized vis-à-vis digital corporations. However, much of this criticism operates within a public law framework as well. Conservative critics also invoke nondiscrimination but shift the focus to who, exactly, is being discriminated against. While early advocates of applying public law concepts to digital corporations warned against the marginalization of minorities or women, others have now flipped the script, often in alarmist terms, to decry the alleged suppression of “conservative” viewpoints or even “masculinity.” These critiques are frequently paired with calls for free speech absolutism and greater majoritarian control over platform governance. Hence, despite their diverging trajectories and rhetorical styles, much of this conservative backlash – particularly its post-pandemic surge – mirrors the public law-inspired argumentative structure of the “liberal” debates from the 2010s and early 2020s. Conservative critics of earlier, more liberal approaches to platform governance now champion a different set of public law concepts. The counterthesis is not that platforms should be governed by rights and adjudication, but rather by majoritarianism and unrestricted speech – both foundational public law principles in their own right. In short, each approach borrows from public law, adapting its principles in ways that are novel, occasionally self-serving, and slightly askew – yet still anchored in a shared frame of reference. This dynamic reveals two key points. First, public law remains the epistemological and discursive framework through which many grapple with new, immensely powerful private institutions like social media companies, and likely others in the future. Second, because of this, the core positions in these debates closely resemble those of classic public law controversies: countermajoritarian difficulty, rights-based approaches, and the legitimacy of institutional constraints.
This book focuses primarily on the first phase of these debates from the 2010s to the early 2020s, a period that might be called the liberal (in the Obama-era understanding of the term) phase of platform governance. That time was shaped by the rise of (or at least discussions about) rights, adjudication, and proceduralism. The book traces how these “liberal” concepts emerged in legislation and managerial practice while also exposing their fault lines, compromises, and contradictions. Crucially, where pertinent, the book also considers whether these approaches should be championed at all or whether more fundamental, structural changes might offer a better path. This question is particularly relevant given the backlash now reaching the highest echelons of policymaking in Washington and Brussels, as well as boardrooms in California – and, increasingly, Texas. Many of today’s critiques respond, at least in part, to the very ambivalences this book uncovers. However, the second wave of public law concepts now reshaping private and regulatory governance – majoritarianism, viewpoint discrimination, and other facets of the conservative backlash – have to be discussed in future research.
The book has an interdisciplinary and comparative approach to connect its findings with related phenomena and perspectives. The book draws from law, sociology, and political science and focuses on two of the three biggest digital economies, Europe and the USA. I dearly hope the book’s findings may be of interest to people working on these issues in law schools, social science departments, bureaucracies, and parliaments, from Brussels to London and Washington, DC, to Addis Abeba, Beijing, Brasília, and Delhi.
One such finding is that building on key elements of what worked in the past to solve contemporary problems inevitably requires careful adaptation. That is especially so because seemingly potent institutions like courts only work if embedded in and enabled by contextual factors. Those contextual factors – the rule of law, widespread commitment to judicial authority, legitimate rules and norms – are, however, seldom reproduced as easily as the institutional shell itself. Simply put, advancing justice and controlling power requires much more than adjudicators and normative material. It requires deeply rooted structural changes in governance, culture, economics, and sociological makeup of the power structure one intends to tame.
Therefore, when it comes to our distinct public–private situation, emulation sometimes carries a whiff of, let us say, moral ambiguity. Clearly, reproducing best practice makes sense. Yet, referencing publicly legitimate role models in private governance – but not their enabling contextual factors – nonetheless goes down so well in general discourse. If there is a flamboyant adjudicator, who cares for the nitty-gritty of practical implementation? In other words, emulation often comes with oratorical grandiosity. However, at times, borrowing public legitimacy by appropriating ceremony and symbolism may not be sufficiently corroborated by contextual factors that enable effectively adjudicating rights in private, profit-oriented contexts.
If that would be it, however, this would be a rather one-sided book. This book does not propose abandoning to emulate well-working older ideas in new contexts. This would be naïve. Learning from what we know makes sense. One cannot reinvent the wheel every single time. Also, not everything that claims to be novel indeed is, hence using what worked well in the past, often, is just fine. And to be fair, even actual courts are far from the powerful change-makers many lawyers would like them to be. So, criticizing novel adjudicators for being not sufficiently effective may be, somewhat, trite. Of course they are not. It took actual courts hundreds of years to incrementally assert their authority – and it remains a constant struggle to maintain that authority.
Therefore, the book also offers a scenario – neutrally, but in my mind not impossible – in which emulated adjudicators potentially do become relevant guardians in the global arena. They do so not because of their formal mandate but because of their public perception and public support. Organizational authority is deeply performative. If you act as if you have authority and people believe you, you kind of have it. From a normative perspective this may be troubling at times. But it seems to be a sociological fact. Borrowing public legitimacy through narration, symbolism, and overall perception potentially empowers new guardians that reproduce existing structures, in our case courts. The two guardians I will study here – Emulated Guardians as I call them – are potentially powerful not because they are courts (they are not) but because they appear like courts. Their organizational setup and normative practice is, too, deeply performative. They seek to control vastly powerful organizations, even though their formal mandate is rather limited. Yet, as they – and this goes especially for the Oversight Board – manage to portray themselves as adjudicators, they tap into the public legitimacy linked to the almost universally acclaimed notion of adjudicating individual rights vis-à-vis a dominant power structure. Surely, now is too early to tell but the performative possibilities that come with emulation may herald a fundamental shift in the governance of private power. Evidently, classic modes of public regulation became – at least partly – ineffective in a globalized and digitized world economy. Yet, demand that power ought to be controlled did not waver. Public discourse still demands accountability for organizations that appear threatening our individual rights. If public regulation cannot do it alone, new actors seem indeed necessary. From a pragmatic perspective, therefore, the actual fight is that for Emulated Guardians’ character and potency, rather than their overall desirability. Emulated Guardians exist and, due to the mechanisms of politics, the information economy and public relations, more will follow.
Building on this sociological rather than traditionally legal perspective, the book eventually argues that the crucial resource of power for Emulated Guardians is not (only) the law but public perception. Effectuating independent, norm-based, ex post control vis-à-vis a dominant new power structure inevitably requires public pressure. Even for Europe, which hails itself for “constitutionalizing” the digital economy, this book shows how many substantive regulatory demands like proportionality and fundamental rights protection will likely remain empty promises if the European public fails to vigorously push for their stringent application in private governance.
In abstract terms, and quite remarkably so, this whole setup reminds of the genesis of judicial review of administrative action itself, which also responded to public demand rather than external regulation. Whenever big changes occur, normativity eventually originates from the people – not from institutions. In the eighteenth century, there was no EU to protect individuals from monarchs. It was the people who did that. The people took to the streets to topple monarchs and fund republics. The people carved out individual liberties from overreaching normative regimes superimposed on them by absolutism and the Church. The people fought for civil rights from abolitionism to suffrage, and various other forms of equality. Surely, some of those fights were supported by institutions like courts. Yet, at the end of the day, holding large, vastly powerful and recalcitrant organizations accountable is a job for all of society, not only lawyers, judges, and regulators. In fact, bereft of societal support, lawyers, judges, and regulators are utterly powerless. Law and regulation do not work in a vacuum. They are a social phenomenon. And if the social current fails to expedite regulatory demands, those demands will become empty ceremonies.
In that sense, highlighting the ambivalence of emulation seeks to challenge the prevailing mood of “depression” among law and technology scholars – a depression that often masquerades as optimism, as Julie Cohen observes.Footnote 2 In their quest for meaning and direction amidst a bleak landscape, many law and technology scholars frequently reinforce, rather than critically examine, self-serving corporate legitimacy claims and the inertia of regulatory complacency. Do not fall for all these performances, Cohen argues in reference to Ari Ezra Waldman.Footnote 3 As discussed in this introduction, I concur with that argument. However, precisely because performativity is such a potent tool – arguably the very foundation of organizational power – it can be directed toward diverse ends. Performativity, discourse, and the social construction of power require a frame of reference. Ironically, in the realm of private governance, an increasingly important frame of reference is public law or, in Cohen’s words, “rule-of-law” language. This dynamic is profoundly influential and could be leveraged to counter the troubling tendencies that Cohen and Waldman so insightfully critique. Therefore, all is not lost. Here, I part ways with some of the more pessimistic voices, as I genuinely believe – and hope this book demonstrates – that there is a positive, even cautiously optimistic, narrative to be found in the performative interplay between public law ideals, innovative lawyering, and private governance; especially so in times of great unraveling and groundbreaking changes. Luckily, Cohen seems to agree, arguing that
“rule of law” language is intended to supply a framework for talking about power, calling it to account, and constraining its systematic abuse. Designing governance institutions for the networked information era requires new thinking about how to translate those broadly framed rule-of-law commitments into mid-level principles capable of being operationalized within networked digital environments.Footnote 4
Thus, in conclusion, the book proposes to think of many of Emulated Guardians and all the other empirical aspects covered here as potential harbingers of a new public law or, not to borrow legitimizing verbiage myself, new accountability norms. These new accountability norms emerge in spheres beyond public authority. They build through public, private, and hybrid forms (these labels are immaterial at the end of the day). Emulated Guardians are crucial interlocutors but only some out of many. Both the norms and the actors shaping this debate remain fiercely contested. The balance between rights-based institutionalism, unrestrained majoritarianism, administrative justice, and entrepreneurial freedom has yet to be struck. After roughly a decade of experimentation with rights, the pendulum appears swinging in another direction since the mid-2020s. In short, these emerging new norms build also on emulations of other forms of control and participation, they are by no means inherently “liberal” but, if at all, in the process of being negotiated. These new accountability norms are effectuated, however, mostly through public pressure and infrastructural control rather than formal legal mandate. This reminds then of the historic emergence of public law itself, which, over a period of hundreds of years, became the mode of societies to control their states. Perhaps, we will see reminiscent processes of – ideally participatory – norm-creation, power-contestation, institution-building, and public participation also vis-à-vis digital corporations (or whatever succeeds them). In this light, today’s struggles, breakthroughs, rituals, and backlashes unfold within a broader historical arc. Nothing is ever entirely lost, and few gains are ever permanent.
1.2 Protagonists
With this prelude complete, I now turn to presenting the book’s two main protagonists: out-of-court dispute settlement bodies (ODSs) and the Oversight Board. Out-of-court dispute settlement bodies are new but ultimately rather weak private adjudicators, foreseen and licensed by the EU to essentially outsource ex-post accountability for content moderation in Europe. They are envisaged to decide millions of cases per year. Out-of-court dispute settlement bodies will be staffed with specialized lawyers and technology experts (although the exact qualifications are unclear) and decide appeals by European users whenever Facebook, X, TikTok, or other social media sites delete or even downrank their communication. However, after much horse-trading in the Council, one of the EU’s main legislative bodies, ODSs decisional authority was gutted, and their decisions are not legally binding. Some hope that public pressure and financial incentives will lend some normative force to the decisions, nonetheless. The hope is that the sheer mass of these cases will improve content moderation in Europe. Another word on vocabulary. There are many ODSs under the DSA, I will speak of ODSs in general here. Whenever I mean a specific ODS, I will highlight that.
The Oversight Board, in contrast, was devised by the highest echelons of Meta, the world’s largest social media company. It is also not a particularly strong adjudicator if one looks at its mandate. It consists of roughly twenty high profile members from all world regions, some former heads of government, journalists, law professors, and a Nobel Peace Prize laureate. Those Board members are supported by a highly capable staff of full-time lawyers. So far, the Board has delivered “decisions” on salient issues like whether Meta could ban then US President Donald Trump from Facebook and Instagram or what type of speech shall be permissible in cases of ongoing violence and ethnic unrest. The Board strategically borrows public law symbolism and personnel, giving normative guidance to Meta on cases it deems representative of structural problems in its governance. This makes the Board a method to channel public critique and orchestrate public pressure vis-à-vis a company that so far appeared largely evasive of comprehensive regulation.
Before we proceed a few words on comparability. As we learned, ODSs are private but licensed by the EU, a public actor. They would not exist if not for Article 21 Digital Services Act, which envisages them as a vital, yet privatized, part of the accountability architecture for online platforms in Europe. The Oversight Board, in contrast, exists because Mark Zuckerberg, Meta’s CEO and dominant shareholder, wants it to. Article 21 DSA emerged in a democratic process; the Oversight Board came about as a management decision. Out-of-court dispute settlement bodies shall, in theory, advance European laws vis-à-vis powerful companies. The Oversight Board in turn was repeatedly described primarily a deflection of Meta’s responsibilities. Does it make sense to compare these two types of bodies?
The answer is yes! This book studies accountability forums of various origin as complementary (and increasingly intertwined) tools to control power. It is not only interesting but imperative to look at public and private approaches together, especially so if the public approach is here to largely privatize a task that can be understood as quintessentially public: individual rights adjudication. Meta’s entirely private approach with the Oversight Board, in turn, takes inspiration from and performatively plays with public conceptions of judicial control over public power. Hence, precisely because the Oversight Board and ODSs are different in many respects it is fascinating to learn how both seek to emulate public law ideas to constrain platforms and, by extension, publicly legitimize platform power. It is their proclivity to emulation (and the desire for the public legitimacy ascribed to it) that is similar – not the emulation’s execution or overall legitimacy from a public law or classic political theory perspective.
Further, things are more complex than schematic public–private distinctions suggest. We must distinguish between public law ideas (like individual rights adjudication) and public actors (like the EU). One key finding of the book is that public law ideas remain strong while public law actors face ever greater challenges to effectively execute their laws vis-à-vis digital corporations. In other words, individual rights adjudication and other public law ideas remain powerful templates to force upon recalcitrant private power structures like platforms. We can learn from what worked to control public power and apply it, mutatis mutandis, to the private power structures that threaten our liberty today. However, it appears, that classic regulation faces its limits when trying to force structurally foreign modes of governance onto profit-oriented actors. Rather, other sources of pressure – public opinion and economics – may be needed.
Equally, we must distinguish between the actors who create the new adjudicators and the new adjudicators themselves. The creators are the EU and Meta, the EU being a public actor and Meta being a private actor. The new adjudicators, however, do not fit as clearly on a purist public/private scale. While the Oversight Board is clearly entirely private in form, it hired mostly people with a human rights background. In turn, ODSs are private bodies but established pursuant to Article 21 of the Digital Services Act. They are licensed and supervised by the EU and would not exist if not for the EU. Yet they are not EU bodies in a formal sense. Simply put, ODSs are privatized and outsourced adjudicators on behalf of the EU. Shuffling a lot of responsibilities on private actors is characteristic for the EU’s approach to platform regulation, and, hence, ODSs epitomize the EU’s current regulatory thrust in the digital sphere.
Both bodies imperfectly mirror the idea of judicial control over administrative action in private governance. Both face the challenge to effectively transfer selected public-law mechanisms into private power structures. Despite differences in functionality and scope, both represent practical evidence for the increasing institutionalization of adjudication-focused public-law norms when it comes to content moderation. In principle, content moderation can be viewed as functionally administrative because it concerns the comprehensive regulation of public rights and infrastructures while also remaining predominantly informed by private law-inspired regulatory forms and, crucially, buttressed by informal, technologically devised, tools and mechanisms. This apparent institutionalization of public-law elements in private governance is a striking development because it incrementally combines North American and European conceptions into a completely new transnational, administrative space dominated by private powers. It is this space where Chapter 6 sketches elements of what could be understood as new accountability norms introduced above. The Oversight Board and the ODSs – which I characterize throughout this book as Emulated Guardians – are supposed to “bring” accountability to that space, protect individual rights, and improve platforms’ internal governance structures.
But the more we focus on this picture, the more the details appear blurred. How far-reaching is the adaptation of public law-inspired mechanisms to private governance? Do glitches occur between the publicly announced and the internally practiced? What are the functional and normative requirements for adjudicating individual rights online – and is it enough to simply increase the number of potential adjudicators? Will those adjudicators’ incomplete institutional design and feeble jurisprudential authority performatively affect our understanding of how “enforceable” individual rights are online? To take the example of privacy, we know how dangerous a continual chipping away of a once venerated institution can be. It has only taken a few years for many – perhaps most – people to accept that egregious breaches of privacy are simply the new normal in the information economy. Will the same thing happen when it comes to other individual rights?
Against this backdrop, this book questions whether Emulated Guardians are, for all their inbuilt flaws, the first significant evidence of the demise of individual rights protection online, or are they potentially the first step toward safeguarding individual rights online? In other words, is public law still the leitmotif to control power beyond public authority – even if not through legislation but more sociological, performative factors? Further, from an EU law perspective, the book offers a fresh, interdisciplinary perspective on the bloc’s attempt to position itself as a global regulator and to – internally and externally – legitimize itself by protecting its citizens from overly powerful platforms. As we will see, the theoretical perspectives and empirical findings offered here paint a complex and nuanced picture of contemporary struggles to maintain legitimate governance and normative orders in today’s increasingly mediated world. Many rules that have been put in place to harmonize the EU’s internal market might end up cleaving normative orders along other fissures. In addition, many of the DSA’s regulatory demands are vaguely worded and, as of now, woefully unspecific; as we will see, this invites a mere ceremonial emulation of public-law mechanisms and ineffective implementation within platforms. If the DSA’s main purpose were to level power imbalances between people and platforms, we can see signs that the EU has unintentionally consolidated platform power because the DSA autonomizes rather than controls how platforms make the rules governing online communication. To end positively though, many of these flaws may be reconciled through subsequent practice, strict interpretation, and, crucially, a vigilant public.
1.3 Contributions
Albeit content moderation is a relatively novel phenomenon, private governance and the EU’s attempts to shape global norms are not. Against the backdrop of seminal accounts on all these issues, this book seeks to make several contributions.
1.3.1 Emulation and Emulated Guardians
First, drawing from organization theory and the political science concept of policy diffusion, I identify emulation as a crucial mechanism for incorporating – or trying to incorporate – public norms into private power structures.Footnote 5 Emulation – in other contexts also called mimesis, imitation, or reproduction – refers to the reproduction of conventional role models in novel contexts.Footnote 6 The key driver behind emulation is the desire for the legitimacy attached to the role model as well as institutional uncertainty over an organization’s future course of action. In principle, emulation is neither good nor bad. However, it often comes with two risks.
Successful emulation requires a contextualized understanding of the problem one intends to solve by it. Emulating governance mechanisms is much more complex than replicating the most visible level of formal structures and institutions. For example, functional – and thereby legitimate – adjudicators are typically embedded within a system of democratic rules and a general culture of accountability. Only the interplay of these factors constructs and reinforces their authority vis-à-vis the power structures they have been constituted to control. On the other hand, if predominantly emulative bodies lack the wherewithal to do the work of control, emulation will often be an incomplete and imperfect phenomenon. Emulation necessitates learning and modification but, once it becomes clear that such emulative accountability structures often lack the equipment necessary to do their job, it may well be merely ceremonial. As anticipated above, I call such novel but emulated accountability structures Emulated Guardians. Given the significance of this phenomenon and the likely increase in demand for similar accountability structures – though perhaps not identical, but certainly analogous – this book deems such a new concept warranted. Whether this conceptual framing constitutes a contribution in itself is for the reader to judge. I would argue it does, as merely labeling these bodies “quasi-courts” or “pseudo-adjudicators” offers little insight into their true nature and fails to reflect their importance.
But why make such a fuss about emulation, especially if it is a standard mechanism of policymaking, institutional development, or for that matter our overall cognitive and behavioral learning? Indeed, we could even dismiss emulative practices as a self-evident and straightforward answer to a particular problem. Yet much legal reform and scholarly commentary repeatedly falls into a common epistemological trap. That trap is to think in dichotomies. Focusing on what I call emulation helps to avoid falling into that trap.
Often, a phenomenon is considered either old or new, legal or illegal, a court or an executive body. Either the Oversight Board or ODSs are perceived as constitutionalizing revolution, or as representing a cabal of corporate reactionism. But emulated institutions, at least in the realm of private power, are more than just shiny updates of old models. Neither should we view them as isolated from their epistemological and sociocultural contexts. In principle, stipulating distinguishable categories is a necessary condition for many arguments, comparisons, and our overall path toward better understanding. For example, this book distinguishes between public and private power, a distinction that, if we are not careful, can end up being decidedly simplistic and even too binary. Contrasting one thing with another can be a valuable way to highlight differences and divergence while also indicating similarities and convergence.Footnote 7 At the same time, we are aware that societal phenomena are rarely as clear-cut as their sociological, philosophical, political, economic, or legal categorizations suggest.
Focusing on emulation therefore highlights, rather than evades the implications of the fluidity of practice – especially in a rapidly developing field such as digitization. Studying phenomena through the prism of emulation – or, as a literary theorist might say, mimesis – means to focus on what a structure or practice picks up from the past and from its context, how a phenomenon digests and newly assembles preexisting ingredients, and to describe the effects this newly fashioned phenomenon has on its context and its own role models.
The power of emulation is to remind us of traditional, revered concepts. Thus, squeezing such phenomena into dichotomous boxes will inevitably under- or overemphasize their potential – as well as their potential dangers. For example, if we compare the Oversight Board or ODSs to traditional courts, it will not take us long to realize that neither of these bodies is a (de facto) court, a human rights tribunal, or any other traditional label that other actors may be trying to stick on them. But does that fact make them irrelevant from a public law perspective? After all, the Oversight Board is an entirely private body and most ODSs will be private as well. Perhaps we should view these as just a particularly sophisticated form of customer service. However, such thinking might prove shortsighted. In a world where private power structures command increasing power over public rights, in this case freedom of speech, calls for accountability of such private power will only grow louder, and rightly so. Yet the resulting struggle for accountability in (privatized) global governance produces phenomena that are inspired by traditional models but hardly ever meet equivalent conditions and contexts in their new field of application. Thus, contrasting the “new” and the “old” risks missing out on the hidden and, as we will see, sometimes rather ostentatious links between then and now, between traditional role model and contemporary emulation. This work, therefore, predominantly focuses on identifying, describing, and contextualizing emulation as a phenomenon in legal practice.
Using the concept of emulation offers a new, critical perspective on many contemporary as well as, arguably, future phenomena in (private) global governance. More power structures will emerge, and more people will come up with ideas for revamping traditional, public-law ideas to control those new power structures. Therefore, instead of concentrating on the alleged “novelty,” “unprecedentedness,” or “innovation” of emergent phenomena, the approach developed here should, I hope, broaden our analytical perspective.
This approach might also make emulation a valuable contribution to administrative law-inspired perspectives on global governance as it unlocks the so far largely concealed door of corporations. Since an emulation-oriented perspective keeps an eye on ceremonial side effects and the risk of appropriation of symbolism through emulation, it also helps in identifying normatively undesirable developments. Focusing on emulation thus enables us weaving together a phenomenon’s past and future, narration, and practice. In this way, focusing on emulation can help overcome such dichotomous thinking.
1.3.2 Performativity
Second, emulating traditional role models for new contexts could have far-reaching potential when it comes to institution-building. Embedded in narratives about freedom of expression, judicial review, and individual agency, Emulated Guardians dip deep into the sociocultural imaginary about protecting individual rights against the state. In other words, the book seeks to highlight the performativity of the phenomena it describes.Footnote 8 These emulative bodies may well eventually modify our traditional concepts and interpretations, especially regarding the enforceability of individual rights online and the accountability of platforms, which – somewhat presumptuously – appear to consider themselves “more like a government than a traditional company,” in Mark Zuckerberg’s words.Footnote 9 Some of those changes might be all to the good, but others perhaps will be worse. For now, the focus here on emulation seeks to mark out the potentially performative effects of Emulated Guardians for content moderation and, by extension, the place of public law and power in a globalized world in general.
1.3.3 Ceremony
Third, once we become aware of the performative effects of Emulated Guardians and the discourse that enables them, the dangers of purely ceremonial emulation become ever more apparent.Footnote 10 Ceremonial emulation presents itself as something (i.e., an effective protector of rights or controller of power) it is not, or at least not to the extent it claims to be. However, by virtue of perceptions of it, ceremonial emulation potentially legitimizes phenomena in the public eye that are scarcely legitimate from a normative perspective. Therefore, as argued elsewhere already,Footnote 11 we should carefully weigh the discursive effects before bestowing symbolism-drenched and potentially legitimizing concepts and roles on these new players. Already at the level of terminology, bodies like the Oversight Board should neither be understood nor narrated using older and sometimes loaded terms like “(supreme) court.” Their essential trait is the fact that they remind us of courts, the procedures of adjudication, and historical examples of the protection of individual rights. In other words, both bodies performatively play with juridical discourse and constitutional metaphors to performatively build up their legitimacy.Footnote 12 To reflect these characteristics, I propose calling bodies like the Oversight Board, or the EU’s out-of-court dispute settlement bodies Emulated Guardians. Constructing such an overarching conceptual frame seems a justifiable way to surmount descriptively vague analogies to courts that might inadvertently help these emulative bodies in being mere legitimizers of the status quo.
1.3.4 Authority and Experiments
Fourth, I explore the practical ability of Emulated Guardians to performatively expand their authority and thereby advance accountability.Footnote 13 Based on internal accounts of EU lawmaking, the Oversight Board’s practice, and the very first, fresh-off-the-shelf ODSs, the book identifies and highlights several inherent shortcomings, unresolved questions, and practical developments to sketch out a reformist roadmap. For example, Emulated Guardians try, with some success, to make up for weak mandates by well-reasoned decisions whose implementation they themselves assess very publicly – irrespective of the fact that the controlled organization is under no legal obligation to abide by the Guardian’s feedback. In turn, analyzing these phenomena through the prism of emulation highlights their reflexive relationship to traditional concepts of individual rights protection, which they partly reproduce but never entirely match. Coming back to the example, Emulated Guardians reference traditional legal concepts, institutions, and styles extensively. They self-present and, arguably, self-perceive as juridifiers of otherwise unruly corporate superstructures. Emulated Guardians’ common frame of reference then is, more often than not, public law. This resemblance without identity might have, at some point, repercussions on the original concept, because the emergence of Emulated Guardians should be interpreted as the next phase in the struggle over individual rights online (and beyond the state in general).
Moreover, since ODSs particularly rely on scaling effects, many are likely to lead the way in incorporating large language models into automated justice architectures. This could provide a valuable opportunity to experiment with such automated systems, especially since these bodies pose tolerable immediate risk – after all, their decisions are not legally binding. It remains to be seen whether this will be a step toward more – and more effective – protection of rights, or whether the path leads toward more ceremonial, and hence less effective, rights protection.
1.3.5 Externality of EU Law
Fifth, from an EU-law perspective, these developments are intriguing as they epitomize the risk that some EU laws may lack normative force at the margins of traditional law.Footnote 14 On the one hand, many hail the EU – and the EU certainly hails itself – as a global regulatory force muscling through European norms and values vis-à-vis other regulators and big business. The Brussels Effect was ubiquitously referenced during the legislative procedures leading to the Regulation (EU) 2022/2065 (Digital Services Act)(DSA) and Regulation (EU) 2022/1925 (Digital Markets Act) (DMA) package. And indeed, with the DSA and DMA the EU has presented perhaps the most ambitious regulation of platform power worldwide. However, the case studies presented below cast doubt on the true normative force these EU regulations will be able to exert. At the moment, the EU’s regulatory approach to content moderation is to infuse principles like objectivity, proportionality, normative clarity, a duty to give reasons, individual rights adjudication into the private governance of platforms. But in its efforts to do this, the DSA is only formulating relatively broad procedural, functional, and normative goals for the platforms. The social media platforms retain “unfettered”Footnote 15 discretion on implementing this normative framework – pushing the door wide open for ceremonial formalities, structures, and procedures that will not settle the deeply embedded governance issues currently faced. Encountering the centripetal forces of private regulation and public pressure, the DSA’s somewhat ceremonial attempt to advance individual rights adjudication therefore autonomizes private rulemaking power instead of confining it.Footnote 16 The DSA has triggered a path dependency that seemingly only has two, equally unsatisfactory, routes. If out-of-court dispute settlement bodies “work” in the way the DSA’s stipulations would seem to be implying, the former may end up remedying only arbitrary rule enforcement while failing to address the deeper problem of unelected, private executives making the rules of free speech in Europe and beyond. If out-of-court dispute settlement bodies do not work, they nonetheless manifest the platforms’ normative force. Out-of-court dispute settlement bodies’ adjudication or inaction will be likely to autonomize rather than legally delineate the appropriate extent of private rulemaking and rule enforcement.
1.3.6 Administrative Perspective on Private Governance
Sixth, regarding content moderation, the book offers administrative law as a complementary perspective to the currently prevailing approaches of constitutionalism, liability, or private governance.Footnote 17 So far, much of the (legal) debate on content moderation has been approached through the lens of (digital) constitutionalismFootnote 18 or viewed through the private-law lens of liability.Footnote 19 Neither of these lenses focuses on internal practices as they are manifested in platforms. Constitutionalism is too loose-knit an approach to fully tease out the technical, but arguably crucial, details of the inner workings of content moderators. In turn, most liability-focused perspectives focus mainly on the external acts of platforms, whose effects might then elicit liability claims, but much less on the internal, structural reasons for such putatively liable acts. Last, some authors are staunch opponents of intensifying government regulation, often deploying variations of arguments that platforms are too big, too globalized, or too innovative to be effectively regulated. This latter approach is, I believe, ahistorical and misguided. Over the centuries we have seen time and again that public law itself is an innovative, flexible, and adaptable mechanism for furthering the control of overweening power. Of course, everything changes all the time, yet few developments are ever truly unprecedented. Platforms create, regulate, and police infrastructures vital to the exercise of public discourse. They regulate – in various ways – the way so many of us communicate online. In short, platforms unilaterally exercise power over individuals in traditional public-law areas, such as public discourse, artistic freedom, the right to do business, and the protection of free speech. Further, regulators increasingly co-opt platforms for policy agendas like the enforcement of criminal law online. Thus, content moderation can be seen, with several caveats of course, as quintessentially administrative. There is a narrative that private power, in our case content moderation, would be so groundbreaking and innovative that “conventional” thinking about guaranteed freedoms and their accompanying guardians would be missing the point. But this story risks chipping away at the basic assumption that domination ought to be normatively justified. Platforms exercise power in crucial public respects. Public law has an orchestra of instruments for controlling power. Within the field of public law, administrative law would offer a much-needed analytical variety to digital constitutionalists, and a much-needed normative program, which would complement still-relevant elements of private regulatory mechanisms like liability. The DSA now ignites another stage of what could be called the administrification of content moderation. The administrative-law perspective proposed here seeks to incorporate and further develop aspects of the Global Administrative Law project and beyond.
1.3.7 New Accountability Norms
Seven, we take a step back and use the empirically validated descriptions of the origins, practice, and potential of Emulated Guardians to reflect more generally on whether the emulation of the specific set of public-law mechanisms, most notably rights and process, which dominate discussions about content moderation (but also about global governance more widely) are indeed the most appropriate normative choice. Perhaps overemphasizing rights and adjudication risks underemphasizing other, more input-oriented, mechanisms like participation in rulemaking. At the end of the day, contemporary reforms of content moderation like the DSA appear incomplete in their focus on the ex-post legitimation of power in individualized relationships. Here, individualized procedures and review mechanisms might help in remedying individual violations. However, many reformers seem to underestimate the fact that most transgressions of platforms vis-à-vis individuals only reflect normative choices made earlier in the process, further up the ladder, and chiseled into platforms’ business models. Reforms thus ought to be much more structural than a mere individualized, ex-post alleviating of symptoms. Therefore, we must broaden the conversation. Especially, there is a much more fundamental discussion to be held on repurposing private law and its mechanisms, not only to incorporate more elements of publicness but, crucially, to reflect the infrastructural realities of technology and code.Footnote 20 However, in these discussions, public law concepts will remain essential. It may be that their efficacy no longer depends solely on traditional regulation, but also on discourse and social pressure. Yet, public law ideas – potentially repackaged and redefined – are such a potent and persuasive “method and mindset” that they will continue to play a significant role in governing private power beyond the state.Footnote 21 In many ways, this book intends to serve merely as a prologue to a much larger conversation.
1.4 Approach and Methodology
This is a socio-legal and a comparative book. It describes and evaluates Emulated Guardians. Both description and evaluation rely on a theoretical frame, and various methods.
1.4.1 Theoretical Approach
The theoretical frame combines three distinct but interrelated strands of literature. These are organization theory, more specifically neo-institutionalist sociology;Footnote 22 the related political science literature on policy diffusion;Footnote 23 and administrative law (in its broadest sense).Footnote 24 Academic works relating to organization theory and policy diffusion have helped to explain why organizations, like the EU or Meta, establish certain formal structures and decide for or against specific policies. In contrast to “rationalist” explanations, which view organizational development or policy diffusion as constituting a “rational” means toward an equally “rational” (i.e. economic, efficiency-oriented) goal, neo-institutionalist and (large parts of) policy diffusion scholarship argue that organizational design, formal structures, and policy or management choices reflect the expectations and “myths” of their institutional environments.Footnote 25 In other words, many formal structures (like novel adjudicators) or policies (such as establishing a new adjudicator and using individual rights as an instrument to control formalized power structures) exist not because they work in a rationalist sense – but because they are socially, culturally, legally, and politically institutionalized and therefore expected. Once institutionalized, a formal structure or policy’s efficacy is hardly questioned, and their “fit” for new contexts is considered natural, obvious, and appropriate.Footnote 26 Sociologists highlighted the role of such “imitation” in the early-twentieth century,Footnote 27 while the first theories of the role of imitation in organizational development emerged in the 1950s.Footnote 28 In the 1970s, tenets of neo-institutionalism began to uproot rationalistFootnote 29 (at least in the economic sense) conceptions of organizational development and the law’s role in it.Footnote 30 However, scholars of administrative law, EU law, and international law have so far not extensively interacted with them.Footnote 31
Emulated Guardians’ underlying idea is to control large-scale, formalized power structures – in particular, social media platforms – by subjecting them to individual rights review by an adjudicator. Social media platforms, partly because of regulatory embedding that the companies which owned them had often unilaterally undertaken, have evolved into huge rulemaking and rule-enforcing structures that can in fact be viewed as “private bureaucracies” (Balkin) and “systems of administration” (Douek).Footnote 32 This functionality mirrors administrative law. The third constitutive strand of literature, therefore, builds on the existing corpus of Global Administrative Law (GAL),Footnote 33 developed since the 2000s and which conceives of administrative law in a broad sense, as pertaining to the theories, doctrines, mechanisms, institutions, and principles regulating the exercise and control of administration-like power.
1.4.2 Interviewing
The genesis of Emulated Guardians is complex, and their practices – in the case of Meta’s Oversight Board which, unlike the EU’s out-of-court dispute settlement bodies, is already up and running for several years at the time of writing – fast-moving and relatively untransparent. To generate insights beyond already published texts and to more effectively interpret the primary and secondary sources at hand, I conducted qualitative, semi-structured interviews with experts over a period of three years. It is essential to emphasize that these interviews served as a complementary source rather than the primary or, in many respects, even a decisive one.
Between September 2021 and November 2024, I carried out twenty-four qualitative interviews with twenty-one interviewees. Most interviews were conducted between September 2021 and November 2022, roughly a third then in 2024 to inquire for follow-ups prior to publication of this book (see Table 1.1). All interviewees were “experts” in that they had special knowledge of the research object due to their occupation.Footnote 34 They represent a diverse set of perspectives from the “inside” and “outside” of Emulated Guardians. Interviewees’ expertise stems from their work in the European Commission, the European Parliament, the Oversight Board, prior or current employment or consultancy activities for Meta/Facebook, newly founded ODSs such as ADROIT, User Rights, and the Appeals Centre, academic work, and activism. Most interviewees had a legal background. They filled positions of varying seniority, with the bulk of perspectives coming from higher mid-level to high-level positions in the respective organizational setup. Most interviewees were embedded in their respective organizations’ hierarchies but still seemed to have considerable leeway in interpreting their roles. Their positions were mostly specialized and professional rather than political or executive. Put simply, in one way or another, most interviewees prepare, influence, and facilitate their organization’s acts but are not the ones publicly accountable for them. The focus on such higher mid-level interviewees was made for three reasons. First, large organizations like the EU or Meta rely on highly skilled staffers.Footnote 35 A great deal of internal wrestling, decision-making, and tacit policy decisions are made by expert staffers and framed in “legal” or “technical” terms.Footnote 36 Such staffers are regularly more entrenched in the details of specific policy proposals, aware of options that did not make it into the published texts, and speak more openly about substance than do PR-savvy executives. Second, analyzing how legal institutions’ “ghostwriters” (in Tommaso Pavone’s term) impact organizational practice is increasingly, and I believe rightly, recognized as a central theme in socio-legal scholarship relating to legal institutions.Footnote 37 Third, higher mid-level staffers are often more accessible than people further up the ladder.
| Date Interview(s) | Interviewee | Institutional Background |
|---|---|---|
| 21 September 2021 | 1 | Trustee Oversight Board, consultant for Facebook during Oversight Board’s creation (2018−2020) |
| 1 October 2021 | 2 | Academic, embedded observer at Facebook during Oversight Board’s creation (2019−2020) |
| 29 October 2021 | 3 | Staffer, European Commission |
| 29 October 2021, 8 July 2022 | 4 | Staffer, European Commission |
| 1 November 2021 | 5 | Academic, consultant for Facebook during Oversight Board’s creation (2018) |
| 17 November 2021 | 6 | Staffer 1, European Parliament |
| 1 December 2021 | 7 | Activist |
| 5 January 2022 | 8 | Staffer, Oversight Board |
| 25 March 2022 | ||
| 13 April 2022 | 9 | Staffer 1, Oversight Board |
| 6 May 2022 | 10 | Staffer 2, Oversight Board |
| 27 May 2022 | 11 | Staffer 3, Oversight Board |
| 2 June 2022 | 12 | Staffer 4, Oversight Board |
| 23 November 2022 | 13 | Staffer 5, Oversight Board |
| 29 June 2022, 16 November 2022 | 14 | Academic, Consultant to the European Commission |
| 15 November 2022 | 15 | Member, Oversight Board |
| 25 November 2022 | 16 | Staffer 2, European Parliament |
| 25 October 2024 | 17 | Co-Founder, User Rights |
| 30 October 2024 | 18 | Founder, ADROIT |
| 30 October 2024 | 19 | Staffer 1, ADROIT |
| 30 October 2024 | 20 | Staffer 2, ADROIT |
| 1 November 2024 | 21 | Director, Appeals Centre |
Generally speaking, the accessibility of interviewees varied widely.Footnote 38 In particular, accommodating confidentiality concerns required lengthy and slow-moving negotiations, a fact that was itself informative regarding organizational habitus.Footnote 39 I retained full control over the research output and confidentiality concerns were met through consent forms and anonymization. Conducting interviews about the DSA faced the additional challenge that the legislative procedure’s speed outpaced expectations, which affected the amount of time when (potential) interviewees from EU institutions were available.
To include a broad set of perspectives, I engaged in so-called snowball sampling.Footnote 40 Of course, the interviewees should not be thought of as a “representative” sample from a quantitative viewpoint.Footnote 41 Instead, the range of empirical insights gathered for this work depicts a plausible account of organizational practices. Diverging accounts of these practices are equally plausible.
Most interviews were conducted via videoconferencing. Except for two group interviews, all the interviews were one-to-one. Typically, interviews lasted roughly an hour. All interviewees gave informed consent to their statements being used in this research. As a principle, interviews were audio recorded and transcribed. Not every interviewee consented to being recorded. For these cases, I made handwritten and computer-typed notes as well as immediately produced “memory minutes.” Unfortunately, this resulted in varying data quality. However, for these exploratory and systematizing interviews, varying data quality can be a nuisance but not an insurmountable obstacle.Footnote 42 All generated textual data (transcripts, notes, memory minutes) were collated and analyzed using MaxQDA software. All data were saved with password-protection locally and on servers located in Germany, and all interviewees have been anonymized. Interviewing was ethically reviewed, approved, and supervised according to the procedures at Humboldt-University of Berlin.
Apart from the interviews listed in the table above, I enjoyed numerous, roughly thirty, “background talks” with activists, academics, EU member-state officials, judges, and other experts. These background talks were not connected to the interview script and did not generate equivalent data (like transcripts, notes, or memory minutes), and are therefore not included as interviews. Nevertheless, they provided valuable contextual material that I was then able to feed into my analysis.
1.4.3 Word Count
The Oversight Board case study contains a brief word count analysis.Footnote 43 The word count analysis quantitatively scrutinizes claims made by interviewees, scholarly commentators, and my own impressions – based on a close reading of all Board decisions issued until November 2024 – that the Board in its normative conceptions focuses mainly on Article 19 of the International Covenant on Civil and Political Rights (ICCPR) and not on the “rules” that Meta’s lawyers and executives have themselves devised. In other words, although the charter and the bylaws envisaged the Oversight Board’s review as based mainly on Meta’s internal rules, the Board arguably turned to external rules to construct a more concise and legitimate normative framework for its adjudication.Footnote 44 To gauge whether the Board was working closely with human rights law on a more general level, I decided to compare how many words the Board allots to human rights analyses in its decisions in contrast to its analyses of other norm-oriented material.
Generating quantitative insights requires an irreducible baseline unit. The Board issues three types of decisions: standard, expedited, and summary. This word count analysis focuses on standard decisions, as they offer the most in-depth normative analysis. All Oversight Board standard decisions (from now on “decisions”) feature similar parts.Footnote 45 Each decision contains a part called “Oversight Board Analysis,” which exclusively contains the Board’s normative reasoning. Initially, that “Oversight Board Analysis” was split into three sections: one section contained the reasoning regarding Meta’s “Community Standards”; another where the case is situated in relation to Meta’s “values”; and one which deals with the particular case’s possible relevance to the provisions of certain pieces of international human rights law, especially Article 19 ICCPR.Footnote 46 In mid-2023, a prior section on Meta’s values was essentially gutted.
The subsection on human rights further subdivides into a three-pronged proportionality test; each prong is announced with a subheading. Counting the number of words that the Oversight Board allots to each of these sections provides such an irreducible baseline unit as it does not require an act of interpretation by the researcher. Each word is one unit, a so-called unigram, and the total number of units in a section – for example, the subsection on human rights – constitutes the word count of that unit. This minimizes the risk of biases entering the dataset. Therefore, assertions regarding word count can be true, whereas more far-reaching claims on the basis of a text’s linguistic or syntactic characteristics rest on complex premises and require interpretation beyond the scope of this book.Footnote 47
A disadvantage of word-counting as a method is that adjudicators, especially with increasing confidence and practice, may tend to express themselves more succinctly or reference an acquis without bothering to explain their normative reasoning all over again.Footnote 48 However, this aspect seems negligible for the Oversight Board as it enjoys discretion in case selection (subject to the prior approval of Meta executives, however, which goes to the heart of the Board’s genuine substantive power).Footnote 49 It seems unlikely that the Board would select cases that fail to raise new normative issues. It does pick cases that cover issues the Board has addressed in the past but which it did not deem to have been sufficiently addressed by Meta – although in these cases the Board typically cites older decisions and presents its argumentation regarding the issue in question.
Based on the interviews, a close reading of the decisions, and existing theories about the value of argumentation and citation in adjudicative texts, I contend that, within its “analysis” section, the Board writes most about what it considers normatively important. Writing more increases word count. Hence, within the Board’s normative reasoning (which is always confined to the “Oversight Board Analysis” part of a decision), each section’s word count may serve as a proxy indicator for that section’s importance.Footnote 50
The Oversight Board publishes all outcomes of procedures (also when there is no decision) on its website. Until the time of this writing in mid-November 2024, the Board issued sixty-eight standard decisions.Footnote 51 All were downloaded as pdfs. As mentioned before, the Board normatively reasons only in the part entitled “Oversight Board Analysis.” Evaluating the total word count of this part and the distribution between and within sections therefore tells us how many words the Oversight Board allots to which issues.Footnote 52 These quantitative insights in turn supplement the close reading and legal analysis of the Board’s argumentation.
All words belonging to the main corpus, headings, and subheadings of each decision’s “Oversight Board Analysis” were manually copied to a table in a Word document.Footnote 53 This table featured one column for each decision and one row for each of the sections (community standards, values, human rights). In the table, the words were manually cleaned of copy/paste errors or hyphenations. Within the rows mirroring the decisions’ human rights section, the sub-sections on legality, legitimate aim, and necessity and proportionality were highlighted in different colors. Any text within the human rights section that did not belong to one of these categories was left unhighlighted.
Using Microsoft Word’s “mark text” and “word count” functions, I counted each section in total as well as the highlighted paragraphs separately. This yielded numerical data relating to the number of words in each section regarding the applicable rules (values, community standards, human rights) and, within the human rights assessment, the highlighted paragraphs (regarding legality, aim, and necessity and proportionality).
Last, I entered this numerical data into an Excel table.Footnote 54 Using basic Excel functions (for instance, column average) and analysis tools (exponential regressions), I then analyzed the numerical data and turned the relevant findings into graphs.Footnote 55
1.5 Situating the Research(er)
Following those methodological remarks, a short reflection on the research’s (and the researcher’s) epistemic background seems warranted. Research is a contingent practice, swayed by discourse, preexisting interests, opinions, experience, attitudinal stance, and temporality.Footnote 56 Generally, findings are contingent on research design, which in turn is contingent on prior epistemologies, interests, and material conditions.
Subjecting powerful organizations to individualized redress resonates with core tenets of contemporary political science, legal scholarship, and culture. Emulated Guardians reflect what their creators thought were effective, legitimate, and legitimizing mechanisms to control platform power (or to be perceived as controllers of this power). But this is only one way of looking at the picture. Reflecting on the researcher’s “subliminal categories of thought” unearths several underlying premises and preconceptions.Footnote 57
As a German-educated (public) lawyer, I tend to have a mild suspicion of anything private that purports to be public. This neither implies a bias against private governance, nor does it reflect a statist agenda.Footnote 58 Martti Koskenniemi considers it “trite” to advocate for the state to “take back control” in transnational governance.Footnote 59 But if private power structures are increasingly exercising such governance, it appears intuitive to install guardians, insert control rods, and contain power. Perhaps the most significant risk in debates about content moderation, technology, and private governance in general is the entrenching of polarized positions between defeatism – “nobody can regulate the Internet” – and simplistic state-oriented approaches – “private governance is inherently problematic.” As this book shows, private and regulatory emulation of public-law legislation manifests itself in an intricate dance of regulators, corporations, public opinion, academics, and activists. Emulated Guardians’ functionality and normative legitimacy does not lend itself to black-and-white thinking.
Another (not so) “subliminal” aspect affecting the form of this book is fieldwork. I conducted numerous interviews with a range of experts from inside and outside the two Emulated Guardians (see Section 1.4.2). The interviews yielded indispensable insights into relatively untransparent institutions. Talking at length with people from various backgrounds challenged some predispositions that I owe to my (German) public-law background.
Therefore, the perspectives reflected in this book echo the diverse voices of the people I talked to. Only their diverse viewpoints enabled a multi-perspectival and balanced analysis. Talking to people crafting the Digital Services Act, working for the Oversight Board or trying to set up an out-of-court dispute settlement body allowed the “research object” to narrate its own version of the story. This is of course never a risk-free undertaking because empirical analysis reflexively influences normative analysis, preconceptions, and understanding.Footnote 60
Meeting the people toiling hard and passionately on these pressing issues inevitably left its mark. Notwithstanding my above-mentioned initial suspicions and the skeptical viewpoint, I sometimes take in the book, this text is certainly not a defeatist critique of lawyers, policymakers, or managers. I was genuinely impressed by the strong commitment of many interviewees to shifting the governance of online discourse onto a better path – despite sometimes crippling institutional constraints.Footnote 61 The culmination of my research and thinking presented here therefore seeks to balance constructive, normatively driven skepticism with empirically informed restraint in light of the complexities of real-world practice. Certainly, engaging in empirical work made me more cautious – the easy answer is often a false friend. Evidently, the observations and arguments presented stake no claims to exclusivity or absolute truth.
Last, a word on AI. As a non-native speaker, my use of artificial intelligence tools has been limited to refining prose here and there. I did not rely on AI in any substantive way beyond improving the wording of text I had already written.
1.6 Terminology and Structure
Before we proceed, a few words on terminology. Throughout the book we will jump back and forth between law and social sciences (mainly sociology and political science), as well as between European and North American perspectives.Footnote 62 This calls for a large degree of terminological caution as not everything that looks or sounds similar denotes similar things in various disciplines or jurisdictions. Thus, before I outline the structure of the book, I briefly highlight what several key concepts are presumed to mean as I use them in each of the ensuing chapters.
In this introduction we have met two dimensions of legitimacy: public and normative legitimacy. Public legitimacy is used here purely descriptively as societal acceptance of a certain power structure paired with the perception that said power structure has the right to act the way it does.Footnote 63 In contrast, normative legitimacy is a reflection of that power structure’s conformity with certain preset norms irrespective of its public reception.Footnote 64 Through emulation, an actor may accrue both types of legitimacy.Footnote 65 However, if an emulation predominantly focuses on establishing publicly visible ceremonies without substantive practical effects, it may yield public legitimacy. This occurs because the ceremonial emulation appears legitimate, as it resembles a normatively legitimate model. Yet, it might also lack actual normative legitimacy, as the ceremonial emulation does not function as effectively as the genuinely normatively legitimate role model. Simply put, constructing something that vaguely reminds of a court, does not immediately make a powerful guardian of individual rights.
Further the book distinguishes between public and private power as well as public and private law. Clearly, ironclad distinctions between public and private law and power would paint an oversimplified picture.Footnote 66 Private law exists within a constitutional framework and is (increasingly) imbued with public-law principles like nondiscrimination. In turn, public law and power co-opt private actors for effective enforcement and reflect, at least in theory and confined by the outer bounds of a constitution, the demands of the “private” majority. Yet “public” and “private” are still useful signifiers that point to significant differences in an actor’s mandate, culture, capacity, and values.Footnote 67 Public, in this conception, refers to organizations associated with state power on the national or international level – or like the EU, for example, which is neither a state nor a typical international organization. Accompanying this observation is the (often implicit) assumption that public actors work, by and large, in the public interest. Private power, in turn, is usually exercised by organizations like companies, which operate on a profit-oriented basis and whose main driver is, in principle, not the public interest but operational benefit.
One last concept that deserves anticipatory attention is institution. In sociological parlance, institutions are stable, repeated, and expected practices.Footnote 68 The term is open-ended and may refer to different phenomena like marriage, parliamentary democracy, or, as in our case, the adjudication of individual rights. In a narrower sense, which is more common in legal literature, the term “institution” refers to a formally constituted organization, like an Emulated Guardian, a specific court or the EU institutions cited in Articles 13 to 19 in the Treaty on European Union (TEU). These are then called “organizations” in sociological parlance.
The book progresses in three steps. The first step, consisting of Chapters 1 to 2, lays the theoretical and contextual framework. After this brief introduction, Chapter 2 theorizes how organizations emulate societal institutions to yield public legitimacy. The second step builds on the theoretical program and presents two empirical case studies of two Emulated Guardians. These are, in Chapter 3, the DSA’s out-of-court dispute settlement bodies and, in Chapter 4, Meta’s Oversight Board. Based on extensive textual analysis and complemented by numerous expert interviews, the case studies lay bare the origins and practice of Emulated Guardians. The third and final step evaluates the empirical findings of both case studies from a public-law perspective. Chapter 5 assesses whether the two Emulated Guardians feature sufficient contextual backing to effectively control platform power and identifies positive, optimistic developments as well as more pessimistic, ceremonial elements. Eventually, both Emulated Guardians feature considerable ceremonial elements but might, in practice, become meaningful accountability forums. For several context-related reasons – especially its turn to international human rights law as external normative framework, its resources, its focus on structural governance issues instead of mass-litigation of individual rights, and its ability to attract public attention – the Oversight Board indeed seems to have taken promising first steps toward providing meaningful accountability. Finally, the last chapter critically reflects on the findings and argues that we witness the emergence of a new set of transnational norms, oscillating between the public and the private. The conclusion, therefore, is cautiously optimistic: while traditional public law actors and regulations alone may not be the answer, public law ideas remain highly relevant in incrementally establishing a form of public guardianship over corporate power structures beyond the state.
Open access