Detecting Propaganda on the Sentence Level during the COVID-19 Pandemic

TL;DR

This study develops a fine-tuned contextualized embedding model trained on Reddit data to detect COVID-19 related propaganda at the sentence level on Twitter, revealing significant differences in activity and content between pro-China and neutral groups.

Contribution

It introduces a novel approach using Reddit-trained embeddings for propaganda detection during COVID-19, highlighting group behavior differences on social media.

Findings

Pro-China group tweeted 35 to 115 times more than neutral groups.

Neutral groups posted more positive and alarmist COVID-19 content.

Pro-China group used more political call-to-action words.

Abstract

The spread of misinformation, conspiracy, and questionable content and information manipulation by foreign adversaries on social media has surged along with the COVID-19 pandemic. Such malicious cyber-enabled actions may cause increasing social polarization, health crises, and property loss. In this paper, using fine-tuned contextualized embedding trained on Reddit, we tackle the detection of the propaganda of such user accounts and their targeted issues on Twitter during March 2020 when the COVID-19 epidemic became recognized as a pandemic. Our result shows that the pro-China group appeared to be tweeting 35 to 115 times more than the neutral group. At the same time, neutral groups were tweeting more positive-attitude content and voicing alarm for the COVID-19 situation. The pro-China group was also using more call-for-action words on political issues not necessarily China-related.