The problem with Meta’s Oversight Board
How should freedom of speech and expression on the internet be handled? The stakes are enormous; there is no forum for speech and expression as public and globalized as the internet. Governmental efforts to regulate speech online have inevitably run into the same set of domestic and international barriers, from simple inability to control online content produced abroad to constitutional limitations.
Which brings us to Meta’s Oversight Board, an initiative by the social media giant that some have likened to a content moderation “supreme court.”
The Oversight Board is an independent trust funded by Meta to render binding decisions on appeals against content removal on Meta’s platforms (including Facebook and Instagram), as well as to make recommendations on broader policy questions. It has been billed as an alternative solution in the absence of governmental regulation on how to handle hate speech, incitement, disinformation and other challenges that arise with online speech.
But the Oversight Board’s handling of a recent policy question, on how to moderate the term “shaheed,” helps illustrate why it is not the answer.
In effect, the Board aims to take questions of great public importance and put them in the hands of a few individuals, largely from ideologically similar backgrounds, in a manner that lacks real transparency and standards. That some members of the board have themselves glorified terrorists only raises further doubts about the board’s moderation recommendations in relation to glorification, legitimization and incitement to terrorism.
• • •
For context, in February 2023 Meta referred to the Oversight Board the question of how to moderate the term “shaheed,” Arabic for “martyr,” given its frequent use as a term of praise for terrorists. Meta presented three possible policy options and asked the Oversight Board to share its views on them. For present purposes, there is no need to get into the substance of the debate, except to note that the Board’s answer was that even the most lenient option — which went so far as to allow for the term to be used in reference to designated terrorists — wasn’t lenient enough.
Shortly after the referral, the Oversight Board invited the public to make written submissions on the question. Over 100 submissions were sent, including one from CAMERA and another from CAMERA Arabic. This process was laudable, although it suffered from some constraints.
Strict length limits meant participants were forced to reduce a complex issue to just a couple of pages. It also meant participants were not given an opportunity to engage with each other’s arguments and pull out their respective strengths and weaknesses. But as long as all those participating were subject to the same limitations it would be hard to criticize the process as unfair.
Unfortunately, there was more.
Unannounced to the public, and unbeknownst to at least some of those who had made written submissions, the Oversight Board also held a series of “stakeholder engagement roundtables,” interactive discussions where participants could more dynamically engage over the pros and cons of the policy options. The existence of these secretive roundtables was only announced publicly after the Board had already made its final decision.
• • •
Shortly after the decision was announced, the Board held a private briefing (held under Chatham House Rules) with those who made public comments. During this briefing, CAMERA inquired as to how the Board decided who to invite to these secretive roundtables and who would be kept in the dark. After all, the roundtable format would enable those stakeholders invited to more robustly and comprehensively engage with the policy questions. Was there a fair process to ensure diverse points of view were included?
The answer, it turns out, was that there was no process. Nameless staff, operating behind the scenes, decided who to invite, and no real explanation was ever given as to how they made their decisions.
This is concerning for several major reasons.
•For one, certain voices are being given an advantage. Worse, we have no reason to believe that those voices were selected for any legitimate reason other than favoritism.
•For another, there’s a lack of transparency regarding an important part of the process whereby the Oversight Board arrives at its decisions. The Board boasts that it has “opened a space for transparent dialogue with the company [Meta] that did not previously exist,” and yet has not shown transparency in its own deliberations and dialogue with stakeholders.
True, transparency is not always realistic. Concerns over a lack of transparency can be alleviated, however, when an organization has established credibility. The public must have a reason to trust in the fairness and professionalism of the process.
Unfortunately, that is missing here, too, in no small part due to the ideological conformity of many of the Board and staff members.
To be clear, many of the Board members and staff are well-intentioned and competent individuals. But what is clear from their backgrounds is that there is an ideological skew, and while there are a handful of ideologically diverse individuals, they appear as a façade of intellectual diversity against a background of ideological homogeneity.
No matter how unimpeachable the integrity and intellect of individuals, echo chambers will negatively affect the quality and perception of their work. This is especially so when that work is to steer how speech is moderated online, affecting an enormously diverse range of voices and perspectives.
• • •
To give some perspective on the bias, one report, by Real Clear Investigations, discovered that 18 of the 20 members of the Oversight Board “collaborated with or are tied to groups that have received funding from George Soros’ Open Society Foundations.” While there is a fair amount of overwrought conspiracy theorizing about Soros’s foundation, it should go without saying that Open Society Foundations is a nakedly partisan foundation that funds projects it believes will advance its ideological goals. That 90% of the Board have some connection to this funding alone suggests a concerning ideological skew.
An open-source review of the backgrounds of members and staff of the Oversight Board further evidences this homogeneity. A handful of activist organizations come up again and again on their LinkedIn profiles, such as the United Nations, the Open Society Foundations, Save the Children, Amnesty International, Human Rights Watch (which has been waging a campaign accusing Meta of “silencing” Palestinian voices via its moderation policies surrounding terrorism), and the Committee to Protect Journalists.
These are organizations which, despite their names and whatever good intentions they may have, are “deeply infected by antiliberal forces,” to borrow the words of one industry veteran. As has been well-documented by CAMERA and other organizations, they are also nakedly partisan and prone to making numerous and often significant errors in the same partisan direction. They are notorious for purveying a highly politicized form of human rights discourse that is often divorced from both common sense and actual international law. In short, the board is full of individuals who hail from a particular “human rights” institutionalist perspective which comes with a host of controversial and ideological baggage.
• • •
Given the secrecy and the partisan bent of the Board, it’s a reasonable assumption that these roundtables were similarly skewed and thus failed to provide the Board with a truly representative, intellectually rigorous debate. What little we know of their nature and format only reinforces this assumption.
In total, five roundtables were held: three “regional roundtables” and two “thematic roundtables.” The three regional roundtables “prioritize[ed] geographies where ‘shaheed’ … [is] commonly used,” resulting in roundtables for the “Southwest Asia and North Africa,” the “Sub-Saharan Africa,” and the “South/Southeast Asia” regions. That is, the Board sought out regions with large Muslim populations, notwithstanding that these regions collectively accounted for just 19% of the written submissions. This is despite Islamist terrorism being a global phenomenon, affecting people far beyond those regions.
There’s also reason to doubt that diverse viewpoints were sought out even within those selective regions. CAMERA Arabic — one of the few Israel-based Zionist organizations with expertise in the Arabic language which had made a submission — was never informed of this roundtable, despite being from Southwest Asia.
Then there were the “thematic roundtables,” one of which perhaps best illustrates the failure: the “Counterterrorism and Human Rights” roundtable. According to the Board itself, they invited a single, unnamed “international civil society organization” to attend.
Few topics are as controversial, including in professional and academic circles, as terrorism and human rights. Yet the Oversight Board sought out only a single organizational perspective, meaning the Board deprived itself of the benefit of the many legitimate alternative perspectives and counterarguments to analyze the issues with various policy options.
Unfortunately, the problems go even further than biases and a lack of transparency and fairness. There are also concerns over what look like potential conflicts of interest and compliance with the organization’s own bylaws.
Nighat Dad, a board member, is also the executive director of the Digital Rights Foundation, which made a submission in line with the Board’s ultimate recommendation. So too did the organization Access Now, which Oversight Board member Ronaldo Lemos had previously served as a board member. In September 2023, board member Afia Asantewaa Asare-Kyei and Deputy Vice President for Content Review and Policy Abigail Bridgman spoke alongside the Electronic Frontier Foundation (EFF) at a conference on “internet freedom.” EFF also made a submission similar in substance to the Board’s ultimate decision.
Whether the Board considered these potential conflicts of interest and took steps to mitigate them is unclear and left unexplained in its public documentation relating to the policy recommendation.
• • •
But perhaps the most alarming evidence of a problem among the board is that some members — who are making recommendations on how to handle content moderation relating to incitement, terrorism and hate speech — have themselves glorified terrorism.
Nighat Dad, for example, has glorified both the notorious antisemite Refaat Alareer and the terrorist Hamza al-Dahdouh. Another board member, Khaled Mansour, has claimed that Hezbollah has fought Israel “heroically” (with a half-hearted hedge “and sometimes … terroristically”) and writes of Palestinian terrorism (“armed resistance,” as he calls it) as mere “details and tactics” that one should “not get bogged down in.”
Helle Thorning-Schmidt, another board member, served as the CEO of Save the Children during a period in which the organization was caught collaborating with kindergartens that held graduation ceremonies that included “mock killing and kidnapping of Israelis by children dressed as combatants.” According to NGO Monitor, they were also collaborating during that period with at least one other organization connected to an internationally designated terrorist organization.
All of this is also on top of the noticeable biases, and lack of credibility, relating more specifically to the Israeli-Palestinian conflict, which — by the board’s own admission — played a central role in their deliberations over the “shaheed” case.
Board members Endy Bayuni and Tawakkol Karman have openly lobbed the objectively absurd accusation of “genocide” against Israel. Karman further claims Israel is engaged in an “unjust aggressive war” in Gaza — notwithstanding it was Hamas who attacked Israel — and celebrates the idea of Israel being a “mere pariah state,” which she accuses of being composed of a “terrorist government” and a “criminal army.” Julie Owono enthusiastically endorses the idea that Wikipedia — a notoriously biased resource which just banned the Jewish civil rights organization, the Anti-Defamation League, from being used as a source of information — is a reliable source of information about the war between Israel and Hamas.
Board member Alan Rusbridger, a former editor at the notorious left-wing and anti-Israel outlet the Guardian, has justified Hamas’s brutal atrocities on Oct. 7 by claiming it “most certainly did not happen in a vacuum.”
• • •
Similar biases can also be found among staff as well, with connections to anti-Israel organizations like those listed above, as well as Islamic Relief Worldwide, Middle East Eye, J Street and the International Commission of Jurists.
Which brings us back to the questions of trust, standards and transparency.
If the Oversight Board wants the trust of the public, and social media companies, to handle difficult questions of content moderation, then it must give us reasons to trust its professionalism, expertise and fairness. The Board has repeatedly criticized Meta for an alleged lack of transparency in its policies around terrorist organizations, and yet the Board itself falls well short of transparency in how it handles its own decision making and dialogue with stakeholders.
Whether the Board came to the right decision in the “shaheed” case is ultimately a matter of some reasonable debate. But the Board’s questionable handling of the public engagement process does not inspire confidence that the issues are being fairly or fully considered.
With upcoming Board recommendations on moderating the phrase “from the river to the sea” and how to handle the use of the term “Zionists” when compared to criminality, it is clear that the Board seeks to have an enormous influence on online expression relating to Israel and antisemitism. As antisemitism and anti-Israel extremism surge, this should be concerning to all.
If the Board wishes to build the credibility it desires, it will need to reconsider its process. Trust is earned, not given. In this context, that trust will require a credible and transparent process.
