Study on Facebook's Oversight Board: Adjusting Expectations

The International Criminal Court is commonly known as the "World Court." With Facebook's announcement that it will create an independent body to review the implementation and enforcement of its terms of service in

The International Criminal Court is commonly known as the “World Court.” With Facebook’s announcement that it will create an independent body to review the application and enforcement of its terms of service among its 2.7 billion users worldwide, the term “World Court” has likely taken on at least one new meaning. Since the creation of this body, there have been voices both for and against it. The points agreed upon by both sides are: 1) that this is a self-regulatory initiative and cannot replace a judicial oversight body; 2) as it currently stands, it addresses only some of the issues users have with content moderation; 3) there are more questions than answers regarding its nature, scope, design, expectations, etc.; 4) the creation of this Board, while an interesting proposal, should not divert attention from due process regarding all content restrictions and curation, transparency in policies and decision-making, as well as the implementation of standards, whether imposed by law or self-regulated. Background and Implications of the Oversight Board After a couple of years of scandal after scandal and a growing set of rules, Facebook is seeking legitimacy to govern what the U.S. Supreme Court has ruled to be “the new public forum.” The rules (http://www.facebook.com/communitystandards) have become increasingly complex and ambiguous over the years, and with each addendum (the internal guidelines are updated roughly every week, resulting in a total of nearly 2,000 changes per year, according to information provided by Facebook). The amount of content “flagged” by users for violating these rules increases significantly each day. Due to pressure expressed in various ways by users and the government, the company has shifted its content moderation from a reactive to a proactive model, has automated most of its practices for detecting unwanted content (though it is unclear what portion of its removal practices is automated), and is facing extremely complex debates on ethics, corporate social responsibility, civil liability, damages, and freedom of expression, among other issues. The decision to create an independent council or oversight board for Facebook, with an open process and consultations, is perhaps the most drastic move we’ve seen from Facebook in recent years, at least with regard to its content policy. And it represents a first step by internet companies toward incorporating external stakeholders into their decision-making processes. After shifting its rhetoric from a focus on freedom of expression to a focus on safety in recent years—and taking on a defensive curatorial role regarding increasingly complex content categories—the Blueprint for Content Governance and Enforcement [NT: the Content Governance and Enforcement Plan] (Nov. 2018) appear to mark a new starting point for the company’s approach to content regulation. First, Zuckerberg’s post directly addresses the need to increase the legitimacy of management and decision-making; second, there is a commitment to create an independent external review board, whose decisions would be binding and public. Many of the details remain unknown. Third, it expressly invites governments to define what they expect from a content moderation regime and details at least two specific initiatives to which Facebook has already committed in Europe: 1) the agreement signed with President Macron to work on new content regulations; 2) its work toward a new European framework for content moderation and regulation over the next two years. Although Facebook had foreshadowed this shift regarding legislation during its 2018 hearings before the U.S. Congress, this is the first official communication Facebook has sent to its own community to welcome—and even compel—governments to regulate. As Zuckerberg states: “Ultimately, services must comply with local laws regarding content, and I believe everyone would benefit from greater clarity on how local governments expect content moderation to work in their countries.” The acknowledgment that “a comprehensive system must address both governance and enforcement” is undoubtedly a positive step forward in the debates on content moderation, particularly those related to transparency. Civil society and academics around the world have expressed concern about the lack of transparency and the levels of discretion enjoyed by internet companies, and have been working with them over the years to improve transparency regarding rules and processes. With this new approach, Facebook can improve the transparency of its procedures and compliance while making binding and public decisions. But perhaps what is even more interesting is the policy shift regarding content management on the platform and what appears to be a new willingness to “share” responsibility (or blame) for rule-making. “As I thought about these content issues, I became convinced that Facebook shouldn’t make so many important decisions about free speech and safety on our own,” Zuckerberg wrote in November. Human rights activists have been saying this for years. What changed? And how much has it changed? The creation of a board that functions as an oversight body for the platform’s content moderation decisions could strengthen the exercise of free speech online just as much as it could undermine it. If the Board is understood as an internal process, designed to help the company address complex issues of free speech and its balance with other rights, unify criteria, and align the system with international human rights standards, the results could be more positive for the entire ecosystem. If it is understood as a replacement for the already weak due-process mechanisms within the platform’s decision-making structure—and as a body that would create and interpret only private law (contract law)—it could harm both the ecosystem and the company. Nor would it fulfill the purpose for which it was created. If what Facebook wants and needs is legitimacy, it can only be built through a dialogue between private self-regulatory standards and existing international human rights standards. Whether or not this Board can achieve legitimacy depends, to a large extent, on how the body is structured, the objectives that are established, the requirements candidates must meet to be part of this body, the nature of its decisions (if it is, in fact, an appeals body), and the transparency and public nature of its decisions and reasoning. Legitimacy in Decision-Making Bodies: The concept of legitimacy can be approached from both a sociological and a normative perspective, and this distinction may be useful in illustrating this point. On the one hand, the more positive the public’s attitude toward an institution’s authority to govern, the greater its popular legitimacy will be. However, this legitimacy is fragile—particularly for the FB Board, which will not have a reserve of legitimacy accumulated over a long history to draw upon. On the other hand, legitimacy can also have a normative foundation, referring to whether authority is grounded in law. Given the global challenge facing Facebook, the legitimacy of the Facebook Oversight Board must be strengthened on both fronts, but even more so on the latter. Building strong normative legitimacy could provide a standard for evaluating the Board and determining whether it deserves support. Furthermore, normative legitimacy can influence sociological legitimacy—or the perception of justified authority—and thus the extent to which it will support or undermine the work of the FB Board. There is a wealth of literature on the elements that contribute to the legitimacy of decision-making bodies. There are even specific works on the legitimacy of international decision-making bodies. What determines their legitimacy? Scholars and practitioners have identified three key elements for the legitimacy of international decision-making mechanisms: 1) fair and impartial decisions; 2) an interpretation and application of rules consistent with their scope and purpose; and 3) that the body be transparent, independent, and governed by democratic norms. Fair and impartial decisions must be at the heart of any decision-making system—not just international ones—in order to achieve legitimacy. Traditionally, “impartiality” has been linked to independence and, therefore, has focused on nomination and selection processes, the quality and soundness of reports, decisions, and recommendations, as well as the public discourse and writings of tribunal members. All these elements have specific and relevant definitions in a vast body of international and comparative jurisprudence that must not be ignored. Fairness, on the other hand, according to Professor N. Grossman, “does not necessarily imply an equal proportion of rulings in favor of and against a particular party; in fact, most human rights adjudicatory mechanisms issue far more rulings against States than against the complainants, and both parties may still consider them legitimate,” but they necessarily require procedural fairness, due process, and a consistent application and interpretation of the law. As for the interpretation and application of the rules in accordance with their scope and purpose, this is likely one of the key elements for the legitimacy of the FB oversight board. Community standards have a scope and a purpose, and they interact daily with other norms, including national, regional, and international human rights standards. Rulings and the reasoning behind decisions in individual cases (if Facebook so decides) must follow the same logic and engage in the same dialogue between community standards and human rights norms. Furthermore, that logic must be transparent and made public. The independent and objective reasoning behind each decision is what a council of experts brings to a closed system such as Facebook’s existing legal team. Finally, transparency in this context can be defined as a quality: an office or agency, whether judicial or otherwise, is transparent “when it creates the conditions that allow society to fully and clearly understand how it operates, the reasons behind its actions, as well as the costs and resources associated with those actions.” This factor also affects the previous two, since without a certain degree of transparency (manifested through any of the following actions: publishing decisions or making them publicly available; providing reasons for decisions; identifying decision-makers, their disagreements and agreements, etc.), there is no way to assess whether a body is biased or not, or whether its decisions and interpretations of the rules fall within the reasonable scope and purpose of the law. While the legitimacy of international decision-making bodies traditionally stems from states’ acceptance of their jurisdiction, the legitimacy of the FB Board would derive from: i) the establishment of a solid foundation of fair decisions; ii) a coherent and persuasive interpretation (persuasion is one of the functions of legitimacy) of the community’s rules in relation to international human rights standards; iii) and holistically transparent mechanisms. Representation vs. Diversity One of the FB Board’s main objectives is to provide greater legitimacy to its content moderation system. Diversity is often cited as a dimension that adds value to an organization’s legitimacy, particularly when the space governed by the organization is diverse. However, diversity has many meanings and should not be confused with representation. Obviously, representation could increase the legitimacy of the board’s decisions in the eyes of the broader community, particularly for those who manage to be represented on the board. However, a 40-member board like the one proposed could hardly represent a community of 2.7 billion members. It is practically impossible to represent everyone. Direct, proportional, or even asymmetric representation could not be achieved on such an unequal scale, and given Facebook’s massive size, there likely could not be a viable ratio to work with. This problem must be acknowledged and incorporated into the design of the oversight board to set realistic expectations. Ignoring it would likely mislead users, creators, and the public and seriously undermine the ultimate goal. In the absence of the representation that a diversity of voices could provide, technical expertise might be the best alternative. It would not be a representative board, but rather a technical body capable of overseeing Facebook’s implementation of its own policy, particularly as it relates to human rights standards. This technical membership should take cultural, geographic, and gender diversity into account in its composition to enable genuine dialogue among different experts and avoid a one-size-fits-all approach to voices and perspectives. Diversity is key to ensuring cultural engagement and understanding across borders; to addressing new cross-jurisdictional issues; and to serving the vast community that Facebook serves. There is no one-size-fits-all model. Deciding what the FB Board will look like necessarily involves defining and agreeing on what it will not be. What is generally expected of the Board? The draft charter suggests that “the Board will be a body of independent experts who will review Facebook’s most challenging content decisions, focusing on significant and disputed cases.” Among the first questions that may arise are: What is the purpose? What is the ultimate goal of having the Board review these decisions? This question directly affects the type of structure Facebook is trying to create. Is this a Supreme Court, an appeals court, a people’s court? Or is it a peer arbitration panel? Could it be an advisory body, or has it already been decided that it will be an appeals body? How will this Board interact with other existing structures within Facebook (Safety and Security, Policy and Outreach, to name a few)? If this is indeed a “Supreme Court-like approach”—which appears to be the basis for its creation—Professors Klonick and Kadri of Yale state that “what this actually means for free speech and due process on the Internet will depend on the answer to one key question: How similar will the ‘Facebook Supreme Court’ be to the Supreme Court of the United States of America?” As they argue in their New York Times article, a key feature of the U.S. Supreme Court (as with any other supreme court) is that it is subject to a set of rules that remain unchanged over time: the Constitution. But, as described in the brief background section of this document, Facebook’s policies change every week, and so far no specific standard or rule has been formally adopted to govern that process. What should be the equivalent for the FB Board? UN Special Rapporteur David Kaye argues that international human rights standards should be the definitive rules governing online content moderation. The adoption of such standards would provide a universal foundation and a shared understanding among many of what freedom of expression means and what safeguards must be considered when limiting it. It would also provide a common language for defining and understanding certain restrictions, avoiding contradictory and ever-expanding terms (such as “hate speech”). Ultimately, the adoption of universal human rights rules—particularly regarding freedom of expression—would also ensure a degree of certainty against discrimination and abuse, whether these originate from governments, users, or advocacy groups, and would help ensure that no one is “a priori” excluded from public debate—a standard established by the Inter-American Court of Human Rights in Advisory Opinion No. 5 of 1995, one of the most progressive and protective international interpretations of freedom of expression and access to information. Treating international human rights standards as “equivalent to the Constitution” has many advantages. However, it also has certain limits that cannot be overlooked. Among these, one of the most important is the platform’s freedom to develop and protect its business and adapt it to different audiences. For example: adult entertainment is not illegal and is protected by international human rights standards. If international human rights law were applied directly, no company could prohibit the distribution or uploading of pornography or violent content on its platforms. However, every company must respect and comply with the UN Guiding Principles on Business and Human Rights, which establish specific expectations regarding the company’s commitment to human rights, its duty to mitigate and prevent human rights violations, its duty to provide redress, and so on. The relationship between international freedom of expression standards and Facebook’s content policies must be promoted, developed, and expanded. Should or could the Council be the body to achieve this objective? The answer to this question will likely help define the Council’s objectives. If it were a decision-making body, what existing models are there, and what elements could be incorporated into this new structure: arbitration systems; judicial courts; international tribunals; media councils; and others? One of the main objectives of Facebook’s oversight board is to lend greater legitimacy to the content moderation system; but creating a body that is sufficiently legitimate for this purpose is one of the main challenges. While this private council to oversee and standardize content moderation decisions is the first of its kind, there are numerous different models that can and should inform the process Facebook is undertaking. First, the company must draw on the best practices and lessons learned from more than 100 years of various types of international decision-making mechanisms if the goal is, in fact, to create a decision-making body; second, these bodies possess a legitimacy that justifies their influence in determining how to move forward with this particular initiative; third, unlike national decision-making bodies, international decision-making systems are specifically designed to address cultural and national differences, having gained experience in this area that should be recognized and from which lessons can be learned. International decision-making mechanisms vary from one area to another and across regions. There are several models that can be considered, including the International Court of Justice; the various arbitration mechanisms established to address bilateral investment treaties (i.e., ICSID); the Human Rights Committee established under the International Covenant on Civil and Political Rights; and the various regional human rights adjudicatory mechanisms (i.e., the European Court of Human Rights, the Inter-American Court of Human Rights, and the African Commission on Human and Peoples’ Rights). These are just a few examples of existing international adjudicatory bodies that have already addressed some of the most challenging issues the FB board will likely face, such as cultural differences, language, nationality, global norms and standards, and international sources of law. The design chosen for the Council will, of course, determine a number of other factors, including the nature and powers of its members, the dynamics expected of the Board, and the means and tools it will have to address diversity in all its forms (criteria, language, legal culture, etc.). Arbitration panels, for example, are case-specific, diverse, and composed of individuals selected from an existing list of accredited arbitrators. They rely on a strong secretariat to keep the process moving and ensure a minimum level of procedural cohesion and some institutional memory that is useful in proceedings that are often confidential. The ability to select arbitrators helps the parties build trust in the process; it provides an opportunity for each party to choose one arbitrator, while the third is generally appointed by the Secretariat. This also helps ensure a certain familiarity among decision-makers with the issues, context, language, and culture in which the case has been presented. The diversity of the panel of arbitrators is particularly relevant to these structures. Even so, the decentralized nature of the body itself can undermine the cohesion of the decisions it produces. The fact that the panel is ad hoc for each case can be problematic when attempting to define, through interpretation, the meaning and implications of a broad or general international norm. International courts such as the Inter-American Court of Human Rights or the European Court of Human Rights are radically different from arbitration panels. They have a stable membership over the years, which ensures a certain degree of consistency and legal certainty regarding decision-making criteria. It also ensures a higher level of equality for its users. Most of the time, there is cohesion in their decisions, and their permanent nature provides greater transparency and accountability on the part of the panel itself. The disadvantage of these structures is that it is not possible to guarantee diversity and representation of all potential parties, and they are likely to be more limited in the number of cases they can handle in a given year. The Inter-American Court, for example, always hears cases in their entirety. The European Court has established a system whereby members of the Court serve on different chambers, each with three judges assigned to it. Some decisions are referred to the Grand Chamber, thereby ensuring broader cohesion in decision-making. As these brief examples show, there are various potential models that FB could adopt when designing its Council. However, regardless of which model is chosen, comparative and past experiences should be consulted and taken into account when finalizing this project. When considering these courts and their practices, special attention should be paid to the lessons learned and the best and worst practices that emerge from them. One of the many problems these tribunals face is the backlog they create. Cases take a long time to be litigated before these bodies, whether or not they require a lawyer, and access to international tribunals is not easy. Among the questions Facebook should ask itself, one of the most important is how this “quasi-private tribunal” will function. Will it add to the systemic crisis worldwide regarding access to justice, or will it help resolve—at least in part—that problem? The goals and expectations for the Council are no small matter to consider: if it is an appeals tribunal—a form of user feedback—how will this body handle its 2.7 billion users and the volume of content they generate? FB already faces criticism for its internal processes for reviewing the company’s decisions on content moderation. This includes not only decisions to remove content, but also decisions to demote or promote the circulation of certain content relative to others. These problems will not be solved by a supervisory board, and, in fact, the creation of a supervisory board should not divert attention from them (regular in-house appeal mechanisms), as they are the foundation for any potential redress for improper or unfair content moderation practices. An approach similar to that of the Supreme Court (or a constitutional or international court) could provide a much more substantial contribution and be more appropriate given the scale of the issue. However, there should be much greater clarity regarding expectations, criteria, the process, case selection, and the board’s stance (in relation to NGOs, users, and consumer organizations), etc., so that this board can receive and select cases, address them, and make them public. The decisions made by the Board should also influence the resolution of similar cases within Facebook’s regular content moderation operations and appeal processes, making the Board an internal reference point and an authoritative body for the company to interpret the Terms of Service in light of international human rights law. Otherwise, due to issues of access, scale, and relevance, the exercise will soon prove irrelevant. Who will serve on this Board? What are the requirements for Board membership? When considering decision-making bodies, who decides is just as important as how they decide. However, our question is not, literally, who will be on the board, but rather what the criteria will be for selecting those members. The question is closely related to the expectations one might have for this board. Who or what is overseeing it, and for what purpose? As explained earlier, selection criteria are essential to ensuring legitimacy. Furthermore, existing peer organizations must be tasked with defining the selection criteria for board members. Knowledge and experience in international human rights—and in freedom of expression in particular—must be among the qualities of any candidate. This recommendation should not be taken lightly. Facebook already has safety and security councils composed of external experts who actively participate in shaping the company’s policies and terms of service. Reviewing decisions regarding content moderation and curation necessarily involves striking a balance between freedom of expression and other rights. Most international human rights decision-making bodies require their members to meet the qualifications for judges in their own countries and to have demonstrated experience in human rights. While not all judges will have the same educational background (some will be from common law countries, others from civil law traditions; some will be from the North, others from the South, according to the North-South divide), all will have some form of legal education; members of the Inter-American Commission or the United Nations Committee do not need to be lawyers, but they do need a certain number of years of experience, a solid knowledge of international human rights law, and high moral and ethical standards. To serve on an arbitration panel, arbitrators must meet certain requirements, including being well-versed in the law. Unlike other structures common to various companies—such as safety and trust committees or child safety groups—the proposed oversight board is intended to address limits and restrictions on speech. While these other bodies are composed primarily of experts on issues related to children, vulnerable populations, risk management, violence, and abuse, it is likely that these groups do not include many experts on freedom of expression. Different areas require different skills, and a key question Facebook must ask itself when defining the criteria for the oversight board is what its role will be and what its skills should be. Conclusions: First, when designing its Oversight Board, Facebook and its team must assess the impact of such a body on the human rights of its users, particularly freedom of expression, due process, access to justice, equality, and non-discrimination. They must also assess how this board, regardless of its structure, will contribute to the implementation of the UN Guiding Principles on Business and Human Rights. Legitimacy is a key element that both FB and its users expect from content moderation. However, legitimacy can be defined in many different ways, and among decision-making bodies, legitimacy requires certain key characteristics that must not be overlooked. The scale and diversity in this particular case—with 2.7 billion users across 180 countries—pose additional and complex challenges to achieving legitimacy within traditional models of representation. These challenges must be recognized and addressed systematically. Facebook should not aim for something it cannot deliver. There continue to be more questions than answers regarding the creation of Facebook’s oversight board. In defining and answering these questions, special attention must be paid to the oversight board’s objectives and expectations. After extensive consultations, it is clear that different organizations and individuals will have different expectations for this body. It is essential to make clear what Facebook’s position is on these points, and it is important to provide new opportunities for dialogue and feedback after the company develops a concrete proposal for its Oversight Board and before it actually implements it. If FB decides to create a private decision-making model, it must take into account best practices and lessons learned from the many and varied structures that decision-making mechanisms have adopted over the years, with mixed results. International decision-making mechanisms are particularly relevant, as they are cross-border in nature, serve a multicultural and interjurisdictional community, function as courts and mechanisms of last resort, and often resolve complex and interrelated legal issues. In accordance with Facebook’s human rights obligations, as well as the specific framework governing restrictions on freedom of expression at the international level, special attention must be paid to the technical expertise required to qualify candidates to serve on the Board. The development of concrete, specific, and clear criteria for selecting Board members is essential for the Board’s establishment, regardless of who selects the first group of members. The selection criteria will depend on the specific objectives and expectations of this Board. About CELE The Center for Studies on Freedom of Expression and Access to Information is a research center located at the University of Palermo in Buenos Aires, Argentina. The Center provides technical legal research to promote the understanding and advancement of freedom of expression and access to information, particularly in Latin America. Since 2012, we have operated the Initiative for Freedom of Expression on the Internet (iLEI), through which we have studied and produced research related to online freedom of expression, access to information, and privacy, particularly within the framework of the inter-American human rights system and standards. Our strategies for promoting change include research; training; and the promotion of high-level forums for reflection and debate. Visit us at www.palermo.edu/cele and at www.observatoriolegislativocele.com. By CELE