The Misty Mountains: Freedom of Expression and the Facebook Oversight Board
In 2018, Mark Zuckerberg announced that Facebook would introduce its own ‘Supreme Court’, an Oversight Board that would be empowered to interpret the scope of ‘acceptable speech’ when users appeal against company’s removals of posted content.[684] In the last couple of years, three moves were made towards making this promise a reality: normative, financial and institutional.
Normatively, a number of documents created the legal framework for the Board, namely its Charter, the agreement establishing a Trust, and the Bylaws and the Rules of Conduct.[685] Financially, an endowment was provided by Facebook as the basis for the already mentioned Trust and an LLC that would ensure the financial independence of the Board.[686] Finally, Facebook initiated and participated in the selection of the Board’s core membership, with twenty members being appointed so far and plans being in place for expansion in the Board’s membership up to forty in total.[687]The Board was introduced into a complex environment, with different states and publics vying over the meaning of speech and facing the dominant role of internally intricate corporations.[688] The Board’s supervisory task was to evaluate speech that is separate from a single overarching legal framework and is enmeshed in an environment regulated by code.[689] The code, in turn, is under sway of powerful international corporations whose economic and technological power[690] often surpasses that of individual nation-states.[691] An additional level of complexity emerges from efforts to regulate this power through legal instruments that do not necessarily forefront freedom of expression, such as data protection and competition law. Social networks in particular have been subjected to both national and transnational attempts to impose innovative but heterogeneous legal frameworks, with somewhat mixed results.[692]
The Board reflects these complexities in its genesis and functions.
While it is a corporate creature, its development was shaped a separation of powers argument.[693] Although created by a corporation, the Board is simultaneously meant to remain independent from the corporate power that holds such powerful sway over online speech. Nevertheless, its introduction may be interpreted as an effort to deflect national attempts to regulate social networks.[694] The Board’s independence is thus a complex undertaking in itself.The chief complexity of the Board and the one we want to focus on here is its relationship to speech. Intended as a moderator of expression on Facebook and Instagram, the Board was originated by a corporation that has not taken a neutral stand to permissible speech in the past. Indeed, Facebook has been positioning itself as an advocate of freedom of expression distinct from any state and their regulatory mechanisms.[695] Its Community Standards have been inspired by the framing of free speech under the First Amendment, but have actually been less permissive than the US Constitution,[696] These Standards make efforts to distance themselves from any national (constitutional) law.[697] They draw on freedom of expression more as the basis for a social imaginary[698] of public authority, rather than a predefined global legal standard.
Klonick helpfully summarises these community standards. Facebook places ‘voice’ as the primary concern, and ‘safety, privacy, authenticity, and dignity’ as secondary values.[699] When translated into the language of freedom of expression, ‘voice’ is intended to be the rule, the space of freedom open for expression of every user.[700] Secondary values are reminiscent of different grounds for restricting freedom of expression. For instance, while a nation-state may provide that speech may be limited if it infringes someone’s reputation, Facebook’s secondary value of privacy allows that content be removed if it deprives others of ‘the freedom to be themselves, and to choose how and when to share on Facebook and to connect more easily’.[701] Some of these grounds limiting voice, however, are more amorphous.
‘Authenticity’, for instance, promises that available content should not be a ‘misrepresentation’, while ‘safety’ allows for removal of content with a ‘potential’ to ‘intimidate, exclude or silence others’.[702] Although Facebook envisions the application of described values in line with international law,[703] in concrete cases their realization does not necessarily have to cohere with any existing standard of freedom of expression.While the Board was expected to restrict its supervision solely to enforcing the Community Standards, its first decisions have taken a different turn. The Board has turned the Standards into a secondary source, interpreting them in light of ‘international human rights standards’ as the supreme source of law. The Board chiefly considers the UN Guiding Principles on Business and Human Rights establishing a voluntary framework for the human rights responsibilities of private businesses, and additionally consults International Covenant on Civil and Political Rights, the General comment No. 34, the Rabat Plan of Action as well as the UN Special Rapporteur on Freedom of Expression Report on Online Hate Speech.[704]
This shift from Facebook’s ‘voice’ to freedom of expression incorporated in ‘international human rights standards’ can be observed in the Board’s first eight decisions in priority cases, which were selected out of more than 220 000 appeals and referrals lodged to the Board to date. With only one dismissal,[705] two rejections[706] and five cases decided in favour of restoring the deleted content, the Board has had an opportunity to define how it understands international standards of freedom of expression and to put them to work on social networks.
The Board has used nearly all of the cases to make procedural and structural suggestions that are not immediately tied to the substance of free speech. For one, the Board found that a case may be heard even if the social network restores a post before its initial removal is placed on the Board’s agenda.[707] Additionally, nearly all final decisions included recommendations for improvement of post moderation.
For instance, the Board found that users should be provided with a reasoned decision on removal and the possibility of appealing automated deletions. It suggested that Instagram’s Community Guidelines and Facebook’s Community Standards be made consistent.[708] It requested that some terms from Community Standards be clarified or, at the very least, substantiated by examples.[709] It called for less intrusive means of enforcing Facebook’s health misinformation policies, especially in light of current COVID-19 pandemic,[710] and argued for additional public criteria and information regarding the scope and enforcement of restrictions on veiled threats.[711]Generally speaking, in terms of freedom of expression, the Board never takes the speaker as an abstract entity floating on the Internet, but as an actor existing as a part of broader social context which should necessarily be considered when making decisions. Independent translation by linguistic experts and research are commissioned to help the Board understand the relevant socio-political and cultural context.[712] Also, public announcements of prioritised case selections are periodically published containing a brief summary of the case, anticipated contentious legal issues and, perhaps most importantly, a call to submission of public comments thereto, from both individuals and organisations feeling they can contribute valuable perspectives that can help with reaching a decision on said cases.[713]
When interpreting the meaning of freedom of expression, the Board either applied established standards or has adapted them. In a group of cases, facts were merely subsumed under established human rights standards to obtain a final holding, such as when Facebook claimed that an at first offensive or shocking statement about a deceased refugee child reached the threshold of hate speech,[714] when an Instagram user practiced her freedom of expression and posted pictures aiming to raise breast cancer awareness[715] or when a Facebook post quoting Joseph Goebbels on the basis of dangerous individuals and organisations rule violation.[716]
Take for instance a case in which Facebook deleted, for violating its policy on hate speech, a post which included two widely shared photographs of a Syrian toddler of Kurdish ethnicity who drowned attempting to reach Europe in September 2015, with a description that ‘[t]hose male Muslims have something wrong in their mindset’.
As Community Standard prohibit generalised statements of inferiority about the mental deficiencies of a group on the basis of their religion, the company removed the post, but the Board concluded that the text is better understood as a commentary on the apparent inconsistency between Muslims’ reactions to events in France and in China. That expression of opinion is protected and, although maybe considered offensive, it did not advocate hatred or intentionally incite any form of imminent harm and as such does not reach the level of hate speech.The other group of cases required new standards or, at least, a significant revision of existing ones in order for them to become more appropriate to their online surroundings. These cases included a video posted on Facebook alleging a scandal at the French agency responsible for regulating health products,[717] and a meme that may be interpreted as a serious threat of violence.[718]
For example, the Board decided a case regarding a post referred to it by Facebook. The Board observed that the user intended to oppose a governmental policy and wanted to change it, referring to a scandal at the French agency responsible for regulating health products. It refused to authorise hydroxychloroquine combined with azithromycin for use against COVID-19, but authorised and promoted remde- sivir usage. As the speaker was making no claims on the existence of a cure nor encouragement to buy and take any drugs without prescription, the Board decided that Facebook had not demonstrated that the post would rise to the level of imminent harm necessary for deletion. The Board nonetheless acknowledged the challenge of addressing the risk of offline harm that can be caused by misinformation about the COVID-19 pandemic. Additionally, since a range of tools exists to deal with misinformation (e.g. providing users with additional context), the company failed to demonstrate why it did not choose a less intrusive option than removing the content as international human rights standards on limiting freedom of expression instruct.
In what follows, we want to place this development in the Board’s approach to its task into a broader context. We want to ask how may this invocation of freedom of expression beyond the quarters of a national constitutional democracy shape the power of private corporations. In the next section, we will present the perspective dominating the emerging literature. Then we will explore what does the constitutional advice paradigm add to the picture.
9.3