Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

SMI-5: Five Dimensions of Social Media Interaction for Platform (De)Centralization

Lynnette Hui Xian Ng, Samantha C. Phillips, Kathleen M. Carley

TL;DR

The paper addresses understanding decentralization in social media beyond network topology by proposing the SMI-5 framework, which describes five interaction dimensions. It develops a case study comparing Twitter (centralized by platform) and Telegram (decentralized per interaction layer) to show how interaction networks—and bot dynamics—vary across views. The key contributions are the SMI-5 framework, multi-view network analysis, and demonstration that decentralization yields distinct metrics and analysis challenges. The findings suggest multi-plane analysis is essential for accurate inference and has implications for platform design, governance, and security in social media ecosystems.

Abstract

Web 3.0 focuses on the decentralization of the internet and creating a system of interconnected and independent computers for improved privacy and security. We extend the idea of the decentralization of the web to the social media space: whereby we ask: in the context of the social media space, what does "decentralization" mean? Does decentralization of social media affect user interactions? We put forth the notion that decentralization in the social media does not solely take place on the physical network level, but can be compartmentalized across the entire social media stack. This paper puts forth SMI-5: the five dimensions of social media interaction for describing the (de)centralization of social platforms. We then illustrate a case study that the user interactions differ based on the slices of the SMI layer analyzed, highlighting the importance of understanding the (de)centralization of social media platforms from an a more encompassing perspective rather than only the physical network.

SMI-5: Five Dimensions of Social Media Interaction for Platform (De)Centralization

TL;DR

The paper addresses understanding decentralization in social media beyond network topology by proposing the SMI-5 framework, which describes five interaction dimensions. It develops a case study comparing Twitter (centralized by platform) and Telegram (decentralized per interaction layer) to show how interaction networks—and bot dynamics—vary across views. The key contributions are the SMI-5 framework, multi-view network analysis, and demonstration that decentralization yields distinct metrics and analysis challenges. The findings suggest multi-plane analysis is essential for accurate inference and has implications for platform design, governance, and security in social media ecosystems.

Abstract

Web 3.0 focuses on the decentralization of the internet and creating a system of interconnected and independent computers for improved privacy and security. We extend the idea of the decentralization of the web to the social media space: whereby we ask: in the context of the social media space, what does "decentralization" mean? Does decentralization of social media affect user interactions? We put forth the notion that decentralization in the social media does not solely take place on the physical network level, but can be compartmentalized across the entire social media stack. This paper puts forth SMI-5: the five dimensions of social media interaction for describing the (de)centralization of social platforms. We then illustrate a case study that the user interactions differ based on the slices of the SMI layer analyzed, highlighting the importance of understanding the (de)centralization of social media platforms from an a more encompassing perspective rather than only the physical network.
Paper Structure (13 sections, 2 figures, 1 table)

This paper contains 13 sections, 2 figures, 1 table.

Table of Contents

  1. Introduction
  2. What is "Decentralized Social Media"
  3. Dimension 1: Platform
  4. Dimension 2: Network Topology
  5. Dimension 3: Content
  6. Dimension 4: Interaction
  7. Dimension 5: User
  8. Differing Views of Interactions
  9. Data
  10. Constructing Interaction Networks
  11. Discussion
  12. Concluding Remarks
  13. Acknowledgments

Figures (2)

  • Figure 1: SMI-5 Five Dimensions of Social Media Interactions for centralized & decentralized layouts.
  • Figure 2: Interaction Network Graphs. Pink represent bot users, green represent human users. Links between two users represent an interaction (e.g., retweet, forward). A centralized platform like Twitter only has a single interaction view, while a decentralized platform has multiple views.