An Analysis of COVID-19 Misinformation on the Telegram Social Network

Abstract

The proliferation of misinformation groups and users on social networks has illustrated the need for targeted misinformation detection, analysis, and countering techniques. For example, in 2018, Twitter disclosed research that identified more than 50,000 malicious accounts linked to foreign-backed agencies that used the social network to spread propaganda and influence voters during the 2016 U.S. presidential election. Twitter also began removing and labeling content as misinformation during the 2020 U.S. election, which led to an influx of users to social networks, such as Telegram. Telegrams dedication to free speech and privacy is an attractive platform for misinformation groups and thus provides a unique opportunity to observe and measure how unabated ideas and sentiments evolve and spread. In this thesis, we create a dataset by crawling channels and groups in Telegram that are centered around COVID-19 and vaccine conversations. For analysis, we first analyze the topics and sentiments of the data using machine learning models. Next, we analyze the time series relationship between sentiment and topic trends. Then, we look for topic relationships by clustering performed on topic-based graph networks. Lastly, we cluster channels using document vectors to identify super-groups of related conversations. We conclude that Telegram communities risk producing echo chamber effects and are potential targets for external actors to embed and grow misinformation without hindrance.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 2022
Accession Number
AD1200512

Entities

People

  • Marcus A. Garcia

Organizations

  • Naval Postgraduate School

Tags

Communities of Interest

  • Biomedical

DTIC Thesaurus Topics

  • Accuracy
  • California
  • Covid-19
  • Data Science
  • Department Of Homeland Security
  • Dimensionality Reduction
  • Disinformation
  • Freedom Of Speech
  • Health Services
  • Machine Learning
  • Medical Personnel
  • Natural Language Processing
  • Propaganda
  • Public Policy
  • Social Media
  • Social Networking Services
  • Social Networks
  • United States
  • Viruses

Fields of Study

  • Computer science

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Cybersecurity.
  • Educational Psychology

Technology Areas

  • AI & ML
  • Biotechnology