I submitted the following public comment to the Oversight Board regarding their review of Facebook’s decision to remove an American user’s Instagram post, which included a picture of Kurdish leader Abdullah Öcalan and text discussing his isolated imprisonment. Facebook justified the removal under its Dangerous Individuals and Organizations policy.

The Oversight Board requested comments addressing “whether the decision consistent with Facebook’s stated values and human rights commitments, including on freedom of expression,” and “the impact of censorship of Kurdish political discussion on social media.” The shadow of the Turkish state looms large in this case – the Turkish government has long conditioned market access on compliance with takedown demands – yet unlimited appeasement undermines Facebook’s commitment to voice as a paramount value.

Oversight Board Public Comment

Will Duffield, Policy Analyst, Cato Institute

The internet may be ephemeral, but Facebook’s offices, servers, and employees exist physically, in the jurisdictions of various governments. Facebook cannot ignore this reality, but its appeasement of governments needn’t be unlimited. American legal prohibitions on Material Support do not extend to third-party praise, requiring coordination or direction to criminalize mere speech. Some geo-blocking may be necessary to maintain access to Turkish markets, but global takedowns are a bridge too far.


Inescapably, to the extent that Facebook’s decisions have real-world consequences, or alter the likelihood of real-world violence, these effects are not one-sided. If the ability to speak on Facebook is important to recruitment and military organization, denying it to the YPG in their multi-front war in northern Syria ensures that more YPG fighters and Kurdish civilians, and perhaps fewer members of the Turkish military, or jihadist groups, will be killed in that conflict.[1]

Facebook’s Community Standards privilege states and government leaders at the expense of stateless peoples. In practice, it leads to interference with the right to self-determination. Article I of the International Covenant on Civil and Political Rights states:

All peoples have the right to self-determination. By virtue of that right they freely determine their political status and freely pursue their economic, social and cultural development.[2]

In conflicts between independence or secessionist movements and existing states, Facebook’s prohibition of the praise or support of “any non-state actor” that “engages in, advocates, or lends substantial support to” “acts of violence” in opposition to a state or with political aims has the effect of capturing even legitimate or constrained nonstate military activity. One need not target civilians to be deemed a terrorist by Facebook. Using these rules, a Facebook operating in May of 1775, reviewing the events at Lexington and Concord, would have deemed America’s founding fathers and Committees of Correspondence “Dangerous Individuals and Organizations”.[3] Should Facebook have prohibited Giuseppe Garibaldi’s worldwide fan following in the 19th century? Even as Facebook banned Tatmadaw accounts in a futile effort to prevent its campaign of ethnic cleaning, any Rohingya who took up arms against their tormentors would be deemed a terrorist under Facebook’s rules.

On the other hand, governments may regularly do violence to civilians, to say nothing of armed secessionists, without risking designation as a Dangerous Organization. In a community of nations loathe to admit new members, some bias is inescapable. There are good reasons to privilege existing governments over revolutionary upstarts. However, Facebook need not pick the government’s side in conflicts as a blanket rule.

To the extent that Facebook makes particular designations of dangerousness, it should make them publicly, accepting comment and criticism. If these decisions are kept secret, they can be made with perfect unaccountability. This might be politically convenient, but it cannot provide legitimate governance. Left unstated, determinations of dangerousness cannot guide user behavior. Secret rules are pretense, not law.

In its January 28th 2020–005-FB-UA Case Decision, the Oversight Board’s recommended that Facebook, “Provide a public list of the organizations and individuals designated as ‘dangerous’ under the Dangerous Individuals and Organizations Community Standard.” In the intervening four months, Facebook has failed to act on this recommendation.[4]

Finally, based on the facts of his circumstances, it seems hard to deem Abdullah Öcalan a dangerous individual. Öcalan is 72 years old and has been imprisoned on an island for twenty years, after being renditioned from Nairobi by the Turkish government. He has advocated a peaceful solution to the Kurdish/​Turkish conflict since his imprisonment. Praise or discussion of Öcalan might, at worst, present an ideological threat to the aims of the Turkish government. Facebook has no business policing such threats.