ALGORITHMIC PERSONALIZATION AND ONLINE RADICALIZATION A MIXED METHODS APPROACH

Abstract

From YouTube functioning as “the great radicalizer” (Tufekci, 2018) to the spread of misinformation on Facebook and Twitter (Allcott, Gentzkow, and Yu, 2019), social media algorithms that personalize user content pose a significant concern for American national security. What users click, like, share, read, watch, and comment on is used by social media platforms as input for their proprietary algorithms that determine future content users see. With those inputs, algorithmic personalization can intensify “reinforcing spirals” (Slater, 2007) in which a user’s media choices affect their interests which, in turn, affect their future media choices and so on, producing a negative feedback loop that can lead toward extremism. Studying the interplay between the technological features of personalization algorithms and the psychological attributes and interpretive strategies of users can help us understand how individuals become radicalized online. Therefore, the proposed research aims to identify technological, psychological, and cultural factors that lead to the radicalization of vulnerable populations on social media through the algorithmic personalization of content.

Document Details

Document Type
DoD Grant Award
Publication Date
Aug 12, 2021
Source ID
FA95502010346

Entities

People

  • Brian Ekdale

Organizations

  • Air Force Office of Scientific Research
  • United States Air Force
  • University of Iowa

Tags

Fields of Study

  • Psychology

Readers

  • Agent-Based Social Robotics and Mobile-Assisted Learning in Virtual Environments.
  • Economics