Federated Social Media

Elon Musk’s takeover of Twitter has reignited debates on content moderation and free speech. But today it is not content that is the challenge but context. We need to move from centralized platforms to federated, protocol-based networks, allowing user-controlled, context-sensitive conversations. This is a model already successful in India’s digital public infrastructure. digital public infrastructure content moderation context data governance.

This article was first published in The Mint. You can read the original at this link.

The news this past week has been all about Elon Musk’s takeover of Twitter and how things are changing in the organisation since he assumed the designation of “Chief Twit”. While much of the discussion has been around the staff layoffs, in the circles I inhabit, equal attention is focused on how his changes will affect the quality of speech on the platform.

Twitter has always seen itself as the world’s public square - the platform everyone goes to for news about events as they unfold. It is one of the few mainstream spaces where people can still speak their mind without fear of being censored - where all the most powerful protest movements of our time - from Arab Spring to Iran - began. It is also the place where some of the most thorny questions around content moderation are getting addressed - what speech should be permitted and on what ground must posts be taken down.

A Wicked Problem

As with all difficult problems there are no easy solutions. Twitter has had its share of controversy - from the suspension of prominent Indian lawyer for posting an antifascist profile picture to the de-platforming of the former president of the United States. It has been taken to task by governments for failing to heed their instructions and by civil society for giving in too easily to their demands.

Musk wants to fix this and make Twitter “the most accurate source of information about the world”. So far, his ideas for how he was going to achieve that objective have ranged from setting up a content moderation council to making the much sought after blue tick available for a price. But I am not sure whether all it takes to solve a problem of this magnitude is just a new hand on the wheel.

Content moderation is a wicked problem. Posts that some see as banal, others view as utterly objectionable. Because these conversations are accessible from everywhere, content that might be acceptable in one part of the world could well be blasphemous elsewhere. No platform can ever hope to moderate content to everyone’s satisfaction - particularly where the issues involved need resolution at global scale.

The Context of Content

The core challenge is context. In an earlier article I had written about how our current crisis of context is making it hard for us to make the most of what technology can give:

We are, today, in the midst of an unprecedented crisis of context. Modern technology has brought more data within our grasp than ever before. However, in doing so, it has failed to provide us with the sort of tools we need to place all that data properly in context.

Content that is innocuous in one context can be deeply offensive in another. Moderation decisions need to appropriately account for nuances across region, language and social group. For a global platform like Twitter, arriving at even the semblance of a balanced outcome in every instance is near impossible.

Platforms have attempted to deal with this by providing content moderators step-by-step instructions as to how these decisions should be taken. They’ve reduced the moderation process to a set of carefully scripted workflows that describe in as much detail as possible what can stay up and what should be taken down. But no amount of detail can ever eliminate the subjective biases that human moderators bring with them. As a result, decisions on functionally similar pieces of content can, often have very different outcomes.

I believe the problem lies, not with the mechanics of the content moderation process, but with the fact that we are attempting to carry out this moderation on a single global platform. By having all our conversations in the same noisy space we’ve increased the odds of conflict between different groups with opposing viewpoints. At the scale at which this is taking place today, it has become impossible for anyone to hope to effectively moderate it.

Federated Chat

To my mind, the answer lies, not in improving the quality of moderation, but in federating our conversations. We need to make it possible for groups of like-minded individuals to organise themselves into dedicated online spaces where they can say what they want without it being taken out of context and causing offence. These spaces could be allowed to connect to each other through interfaces that users can adjust so that they only get to hear and see content they want based on their own personal tolerance for different types of speech. If we can do this we will move decision-making away from the centre and out to the edges of the network. In doing so we can ensure that power is no longer concentrated in the hands of a few decision makers in large tech companies.

To implement this we will need to make a shift from platforms to protocols. We will have to renounce our dependence on platforms for our social networking needs and instead embrace federated social networking based on protocols that offer the freedom to create content and interact with others in a manner that we can more fully control.

Protocols not Platforms

This is not as daunting as it might sound. For a while now, federated networks like Mastodon have provided their small but growing user base a calmer, less frenetic alternative to existing social networking platforms. Others like Blue Sky are in the works. These new environments take some getting used to - particularly for those of us accustomed to the dopamine-inducing algorithms that power the popular platforms. But once we understand how they work, the freedom and flexibility they offer are refreshingly different.

None of this should feel strange to us in India. We have, for over a decade now, eschewed platforms in favour of protocols. All our digital public infrastructure - from UPI to DEPA and Beckn - has been built in this manner, demonstrating how diverse digital systems can connect to each other to share data while still enabling private innovation and customer-focussed innovation.

It is time to take the lessons we’ve learned and apply them to social networks.