[Unfinished Thoughts] Content Moderation, International Relations and Ideology

I have recently given a guest lecture for International Relations students at Tecnológico de Monterrey, Mexico. For it, I wanted to make sure to link some of the classic IR literature to the issues of content mod on social media. I succeeded and failed. The talk ended up being about underlying ideology and politics, which made for great discussion.

I want to use this space to share some of the more crucial elements of the presentation for the sake of public profit. Also, this takes the form of a massive braindump which represents my own views and no one else’s. And thus will be subject to revision without notice.

The current elements of the social media discussion

To start, we need to ask ourselves how we got here and what are the core elements that permeate the social media discussions. In short some of the elements are: (1) the nature of the internet; (2) the replication of human behavior – good, bad and ugly – in that hyperreality; (3) each country’s freedom of expression tradition; and (4) an increasingly stingy media ecosystem that lives from controversy and theatrics instead of helping understand the underlying nature of social media companies’ way of working.

In that environment it is important to keep Hanlon’s razor in mind when thinking about social media company behavior: “Never attribute to malice that which can be adequately explained by stupidity.” Mostly because I believe that inventing motives distracts us from thinking more deeply and critically about the issues at hand.

This being said the following words will be directed mostly shedding some light into the nature of State-Transnational Company relationship and what are some of the ideologies that support the structure of some of the most talked about governance ideas. Two main ideas:

  1. The problem of content moderation and platform governance/regulation requires a politico-ideological solution.
  2. This is the case because the issue at hand is a wicked problem which technocracy/technology/engineering can only partially solve.
  3. What is worse, if we continue to be inclined to leave the solutions to engineers we run the risk of not being able to influence the ideological path dependency that some of these solutions will create.

The International Landscape that supports social media companies

Social Networks and global internet actors flourish in a very “neoliberal” view of international relations. In it, cooperation and transnationalism are viewed as keys to integration, leading to better economic and political outcomes (e.g. less chances of war). In a neoliberal view of international relations, companies, engaging in Surveillance Capitalism or otherwise pushing for more classical capitalist globalization, stress the classical concept of Westphalian state sovereignty. In short, power that was previously circumscribed within geographical borders is relocated to private entities ruled by incentives that differ drastically, for the most part, with those of a classical nation state.

In fact, if we take Bull’s ideas of the “anarchical society” we will find that he postulates that States organize internationally in a society. This society is held together by a sense of common interests and ideals (e.g. human rights). These states also build common rules and institutions which, in turn, creates a larger sense of belonging. However, international society is anarchical (i.e. there is no state that solely holds the power to enforce rules over the others. In other words no one holds the absolute monopoly of force). And in that context, as Wendt said “anarchy is what states make of it”. Or, in this scenario, anarchy is what social media companies make of it.

In this context there’s some that believe that the existence of an international society (and the peace it consequently generates) is reinforced by transnational economic ties. Which would make them great proponents of the continuous existence of transnational actors that erode the State’s traditional power.

Interestingly, the question of digital sovereignty and of transnational powers syphoning power from our classical nation-states is not a new debate at all. Since the late 90s, beginning of the aughts, there has been scholarly debate around these issues. In it some relevant questions have appeared: (1) what power do States hold in an anarchic world where services aren’t necessarily bound to local standards because of their global, self-regulating, nature? (2) What does sovereignty mean when we speak about social networks and cyberspace at large?

What does this international landscape mean for the “global south”

The “centre vs. periphery” logic that Raul Prebisch used to understand the differences in power and economic development between countries seems to be a great analog to what is happening in terms of digital regulation. The balance of power is inclined in favor of countries/regions that have a higher economic power or happen to be the host/headquarters of social media companies.

This fact, creates an international power imbalance by which countries in the global south are submitted to regulation/legislation that comes from other States. The main question here is what our unit of analysis should be. Some legislation comes from country/regional blocs versus other that comes from states that are part of a larger federation (e.g. California). The locus of power matters because it resolutions affect people across the globe as they force (or set incentives) transnational companies to behave differently across borders. In fact, the question of locus of power might be more complicated because we need to take into account who makes the decisions, who writes the standards and who is affected by them. Most of the time these three sub-sets of people do not overlap.

This situation begs the question of what countries in the global south can/should do to balance International power vis-à-vis other states and social media companies. In other words: What power do States in the periphery hold, in an anarchic system, in which social media services are not constrained by local standards but by either auto-designated standards or by standards required by foreign states?

To build on top of this question, we need to wrestle with the fact that the technical aspects of content moderation (economies/problems of scale, automation, etc.) have a direct influence on the realm of the socio-political. However, the inverse relationship does not happen as easily. Companies, specially policy and engineering teams hold the power of defining the structures of content moderation systems – and those are part of the companies priorities, not the users.

Content Moderation Governance: Powers in tension

At a very high level, there seem to be three main ideological powers that are fighting for contention. The way that States/companies have been trying to organize themselves can be characterized in two buckets: the globalists, the nationalists, (and the self-regulatory).

As stated above there are globalizing forces at play that we need to be mindful of. Grosso modo I have identified four elements – not exhaustively -that play in favor of globalization: (1) the decentralized and geographically unbounded original nature of cyberspace; (2) the minimum standards of Human Rights and international law that ensure freedom of expression and strict tests based in jurisprudence to censor speech; (3) the idea of content moderation “cartels” coined by Evelyn Douek, whereby platforms tend to make the same content decisions on high-visibility issues because of media ecosystem pressure; (4) the idea that content moderation is “assembled” (in the sense that Beck uses the word in his book “Risk Society”) over infrastructure, and therefore elements at that underlying level trickle up towards the surface layers (content) in terms of what is allowed to exist (payment platforms, data storage, etc.). Examples of this last phenomenon are payment restrictions for pornography or even rules on what is allowed to be hosted in certain servers.

On the other hand, there are “nationalistic” forces. Several States have tried to fight back against the power of transnationalism and they have done so trying solutions at three different levels: systemic, procedural and theme-based. Examples of systemic approaches are the creation of “state-intranets” like in North Korea and Iran where the power is derived from controlling the whole network. A level lower, is the idea of controlling procedures of content moderation which is what Germany does with NetzDG and France tried to do with its failed Avia Law. The idea that companies needed to adapt reporting and reviewing flows and turn-around-times to adapt to these laws shows the power of them – although their actual effect is unclear in terms of whether they actually force different outcomes altogether or if they offer heightened protection for their citizens. Finally, the idea of theme-based legislation is what most countries do by default because of their application of older rules to the content moderation space. A classic example of this is the regulation around the topic of defamation.

The main issue with nationalistic tendencies is that they are, at their core, opposed to the universal nature of the internet and its services/networks. Therefore the more enforcement at country level we see, the larger the risk of “balkanization” of the internet. Where we lose all the benefits of networked connection because platforms aren’t able to operate across borders because of legal risks.

Finally, there is a “third way” which I’d characterize as platform “self-regulation”. This would differ from their globalist power in the sense that it requires them to create new institutions that are independent and relocate some of their central power to them. This is the case of Facebook’s Oversight Board. Some of the main issues with the creation of this type of organism, at an international level, is that a private company continues to relegate power in a private structure that is outside of international agreements (even though in the case of Facebook’s Oversight Board one of the core elements of it is their tie to human rights law) but most importantly, which States do not necessarily have any power over. In this scenario, democratic power whether of countries or users is not necessarily reflected properly.

Moreover, self-regulating bodies in content standards seem still to be far away from being a silver bullet. Cases presented to them are never straightforward to “solve”- as they are supposed to see the most relevant ones –, but most importantly it doesn’t solve the problem of local tolerances being heterogeneous which means locally made decisions might differ with the originally globalist idea of procedural/decision fairness (“equality under the law”). In other words, what is behind this is, how do we conciliate global policies with local standards? And, most importantly, who and how do they make decisions? Self-regulating bodies seem to be a good step in divesting power out of private companies and into bodies with different incentives but they seem to be subservient to the originally globalist platforms.

Ideological reflections

The question of what kind of content moderation model we want to have is a political philosophy question. Within this realm there are interesting defenses of democratic pluralism versus technocratic/neoliberal solutions. First of all there are substantial and logical arguments in favor of embracing the heterogeneity of communities that use the service by decentralizing decision-making power. First and foremost there is the benefit of increased information transparency. As I have argued in past writings the current social media ecosystem does not foster decision-making level intelligence for users or regulators (in fact, most of the time is the complete opposite). Second there would be higher levels of accountability at user and central level because more scrutiny, checks and balances, would be placed on both ends of the institutional design. Third, decentralization means that we’re able to replicate the institutions recursively to the level of granularity needed. Fourth, there’s an argument of cost-effectiveness at the central level – which we can delve more into when we speak about economies of scale. Fifth, it satisfies those who believe in the “Leviathan hypothesis” in which the State is a greedy institution that pushes for infinite resource consumption and growth.

However, decentralization and democratic pluralism comes with its own set of complications. At a high level, accepting content moderation decision-making alternatives based in democratic pluralism (incl. decentralization) that emanate from private companies means to ratify their power as global actors. And this begs the question whether while we’re trying to push and follow an ideal of user (“citizen”) participation we are concurrently de facto ratifying these companies as pseudo/neo-States?

In other words, does pushing for more democratic social media platforms, mean that we’re cementing these extra-State organizations whose main reason for existence is to continually produce higher economic profits? It would seem that in this situation the question of the difference in leitmotivs between a nation-state and a company is critical. While a nation-state, for the most part, has the goal of serving and governing its people and territory, a company’s existence is determined by serving its shareholders who, at least in theory, just want better return on investment. This is not to say that sometimes the idea of generating more revenue cannot be aligned with what would be “best” for the users. What is really clear, is that it that “sometimes” is not nearly enough.

Final Thoughts

Power is not equitably distributed at an international level. When thinking about new actors in international relations we need to ask who has veto powers and who are the de facto powers.

Simultaneously, we need to start doing more work to unveil what the implicit ideologies are. This will become critical to understand who’s interest it is for a specific governance model to reign.

That being said, there are no perfect solutions or silver bullets. This means that our choice will and has always been ideological: just like we defend democracy as our governance “north” what is that equivalent for content moderation or even the meta-verse?

See you in the next adventure,



Bonus: International Relation Concepts

Below I list some of the relevant I.R. concepts that I think are useful when thinking of social media governance and regulation: