What the United Nations Guiding Principles for Business and Human Rights (Don’t) Say About Content Moderation
In 2020, one of the Facebook Oversight Board’s first cases dealt with Facebook’s decision to delete a post alleging that a cure for COVID-19 existed. The board decided that Facebook ought to reinstate the comment because its deletion was an unproportional measure under the test set forth in Article 19 of the International Covenant on Civil and Political Rights to evaluate restrictions on freedom of expression. This is just one example of how the board uses international human rights law (IHRL) as its own framework to justify all of its decisions. This framework conveys the message that the board’s members are not merely an editorial board issuing decisions and recommendations based on their normative taste. Instead, the board presents itself as an executor of the global public interest, purportedly embedded in international law. Just like courts articulate their policy preferences as legal reasoning, the board tries to persuade us that its decisions are an application of exogenous principles with global legitimacy.
But are the Oversight Board’s decisions really an objective application of widely accepted international rules? In a forthcoming paper, I argue that this is not the case. The board uses IHRL in an imaginative and incoherent fashion, often widely reinterpreting international law principles and instruments. Thus, even if IHRL reflected some global agreements, it is doubtful that the board’s use of these norms and instruments reflects that consensus. This is bad not only because it is misleading as to what international law requires but also because it conceals the need to figure out what the actual global public interest is in platform-moderation policies.
The board’s decision to apply IHRL in the first place is a good example of the board’s idiosyncratic interpretations of international instruments. The board claims that its application of IHRL is compelled by the United Nations Guiding Principles on Business and Human Rights (UNGPs). The UNGPs, which were endorsed by the U.N. Human Rights Council in 2011, are a set of guidelines for states and companies to prevent and address the human rights abuses committed in business operations. Adhering to an increasingly popular interpretation of the UNGPs, the board states that it follows IHRL per the UNGPs. However, a close look at the text, history, and different principles of the UNGPs shows that this is a misleading reading of the instrument.
The Board’s Reading of the UNGPs Contradicts the UNGPs’ Text
The UNGPs expect companies to take measures to stop any contribution to human rights violations. In that sense, they do provide guidelines that would help reduce the harm caused by online speech. This responsibility means, for instance, that companies ought to take action regarding speech that international law prohibits.
International law, however, does not currently recognize a human right to speak on privately owned media without being subject to the medium’s editorial policies. Neither does it establish a corresponding duty for media owners to host all speech that international law protects. On the contrary, as Molly Land and Tarleton Gillespie have argued, platforms are valuable to users precisely because of the choices they make about what content is available and how it is organized. Accordingly, because excluding lawful content is an essential aspect of what social media companies do, IHRL does not necessarily fit well with how platforms operate. In this context, it is unclear which human rights Facebook or its Oversight Board is expected to respect.
The Board’s Reading of the UNGPs Departs From the UNGPs’ History
The history of the UNGPs shows that the board’s interpretation does not follow global agreements. Previous U.N. attempts to pass guidelines for business and human rights that aligned with the board’s interpretation of the UNGPs were explicitly rejected by the Human Rights Council.
Since the 1970s, the U.N. has tried to address the negative impact that corporations can have on human rights. The direct antecedent of the UNGPs were the failed U.N. Norms on the Responsibilities of Transnational Corporations and Other Business Enterprises With Regard to Human Rights 2003 (Draft U.N. Norms). The Draft U.N. Norms imposed on businesses the same general obligations of states to promote, secure the fulfillment of, respect, and protect human rights under the human rights treaties that states have ratified. The project failed because the corporate sector and most states believed that companies should not implement international law themselves.
In that context, the Human Rights Council created international relations scholar John Ruggie’s mandate in 2005 to draft new principles. Ruggie decided to avoid the debate that had bogged down the U.N. Draft Norms. Thus, he focused on distinguishing states’ and corporate obligations. To begin with, states have duties while companies only have responsibilities. But the degree of commitment is not the only difference between states and firms. Rather, the content of the commitment is different. According to Ruggie’s framework, companies are not expected to implement international treaties. They should look at human rights treaties not as a source of rules that apply to them but as an enumeration of recognized rights they should not infringe.
For example, imagine that a state informally asks a cybersecurity firm not to serve a specific client because the state disfavors the client’s political views. The client has a human right not to be silenced by the state. The UNGPs set the expectation that the company will respect that human right and will not contribute to its violation regardless of the company’s domestic legal obligations or lack thereof. The UNGPs, however, do not set the expectation that the company will make decisions about which clients to serve as if it was a state, that is, by implementing the same principles that states must follow when regulating expression.
The board interprets the UNGPs in a new fashion. It claims that Facebook should look at the text of international treaties—and the interpretations that the U.N. Human Rights Committee and U.N. special rapporteurs have made of them—as a set of rules that the company itself should apply on its platforms. This is the approach adopted by the Draft U.N. Norms, which the Human Rights Council rejected and from which the UNGPs explicitly diverged.
The Board’s Reading of the UNGPs Makes the Principles Incoherent
The board’s interpretation of the UNGPs also alters the interplay between states’ duties and corporate responsibilities. Arguing that every content moderation decision implies an IHRL issue means that states have a corresponding obligation to regulate each instance of content moderation. The board’s interpretation of the UNGPs assumes a human right to freedom of expression on social media platforms vis-a-vis the owners of such platforms. If individuals have that right, states have a duty to protect it. However, there is little support, if any, for the idea that states must protect the exercise of all IHRL-protected expression on private platforms. Indeed, no state requires companies to host all speech that international law protects. Ultimately, the board reads the UNGPs as setting the expectation that companies will afford a level of protection of freedom of expression stronger than what international law requires from states.
The Implications of the Board’s Reading of the UNGPs
By linking its decisions to international law, the board implies that a global agreement on how platforms ought to moderate content exists, when it does not. Indeed, the board claims to be an enforcer of preexisting global standards. In the example examined here, the board construes the UNGPs as an instrument that sets the expectation that IHRL will be the framework for moderating content on social media platforms. That is a significant and nonobvious normative preference (especially given that international law arguably protects the denial of mass atrocities, blackface, racial slurs, and other discriminatory speech that Facebook prohibits). But the board does not say that international law aligns well with its editorial taste. It tells us that its reliance on IHRL results from the direct application of a U.N. instrument.
Ultimately, even if IHRL or the UNGPs reflect global values and commitments, the board’s interpretation is so transformative of those principles that its use legitimizes the board’s decisions without necessarily advancing a preexisting global agreement.
More importantly, the board’s use of the UNGPs illustrates the main risk of the board’s resort to international law more generally. The board’s performance as an executor of international law rests on the assumption that IHRL can function as an account of the global public interest, capable of producing rules and decisions that also reflect those values. But this interpretative effort conceals the need to figure out collectively what the public interest is. The risk is that the board undermines its most laudable purpose. While one of the board’s main contributions is to include the public interest in social media governance, the board’s use of IHRL suggests that the public’s participation is actually unnecessary because its interests are already embedded in IHRL.