Zarine Kharazian
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Zarine Kharazian has a total of 11 co-authors. Here are the top co-authors by number of shared publications:
- Benjamin Mako Hill
University of Washington
2 shared publications - Kate Starbird
Human Centered Design & Engineering, University of Washington, Seattle, Washington, USA
2 shared publications - Carly Minow
University of Washington, Seattle, WA, USA
1 shared publication - Jevin D. West
University of Washington
1 shared publication - Kayla Duskin
University of Washington Information School, University of Washington Center for an Informed Public
1 shared publication - Lukas Erhard
Department of Computational Social Science, University of Stuttgart, Germany
1 shared publication
Latest Publications
- Open Access
- 2 May 2026
- Published by Center for Open Science in SocArXiv
Epistemic Infrastructure Under Attack: Conceptualizing the Capture and Destruction of Public Knowledge Institutions
From organized disinformation campaigns on social media to politically motivatedattempts to take over publishers and platforms, the last decade has seen a proliferationof attacks on the epistemic infrastructure that people rely on to understand the world. Weargue that scholars, practitioners, and policymakers have struggled to respond effectively tothese threats in part because they lack systematic ways to understand the shifting nature ofwhat is under threat and how it is being attacked. Drawing on Elinor Ostrom’s classic workon commons governance, we develop a conceptual framework that conceptualizes epistemicinfrastructure as a shared resource system whose three interdependent components—knowledge, governance, and legitimacy—can each be attacked independently by motivatedactors through two distinct modes: capture and destruction. Applied to recent attacks onWikipedia, public broadcasters, and government data agencies, the framework reveals howattacks frequently target governance structures and institutional legitimacy rather than theknowledge outputs and information flows that existing approaches focus on, with implicationsfor how scholars and practitioners diagnose and respond to these threats.
- 2 May 2026
- Published by Center for Open Science in SocArXiv
Open AccessEpistemic Infrastructure Under Attack: Conceptualizing the Capture and Destruction of Public Knowledge Institutions
From organized disinformation campaigns on social media to politically motivatedattempts to take over publishers and platforms, the last decade has seen a proliferationof attacks on the epistemic infrastructure that people rely on to understand the world. Weargue that scholars, practitioners, and policymakers have struggled to respond effectively tothese threats in part because they lack systematic ways to understand the shifting nature ofwhat is under threat and how it is being attacked. Drawing on Elinor Ostrom’s classic workon commons governance, we develop a conceptual framework that conceptualizes epistemicinfrastructure as a shared resource system whose three interdependent components—knowledge, governance, and legitimacy—can each be attacked independently by motivatedactors through two distinct modes: capture and destruction. Applied to recent attacks onWikipedia, public broadcasters, and government data agencies, the framework reveals howattacks frequently target governance structures and institutional legitimacy rather than theknowledge outputs and information flows that existing approaches focus on, with implicationsfor how scholars and practitioners diagnose and respond to these threats.
- Open Access
- 8 February 2026
- Published by Springer Nature
- p. 83-104
The Anatomy of Rabbit Holes: Studying Information Segregation in YouTube’s Recommendation Graph
The impact of YouTube’s recommendation algorithm has been subject to much debate. Case studies and anecdotal evidence have suggested that algorithmic recommendations may lure the platform’s users into informational ‘rabbit holes’ where they are exposed to biased content and misinformation. However, empirical research on the phenomenon has been challenging to conduct because the effects of the algorithm are inherently confounded by user and creator behavior. In this chapter, we study the net effect of algorithmic recommendations through a comparative network approach. That is, we collect the recommendation links between videos for various political and non-political issues. The resulting networks are then compared on various structural network characteristics. If the algorithm creates rabbit holes, one would expect the recommendation networks of videos on conspiratorial content to be more modular, less centralized and intra-topically dense. We analyzed 15,455 videos and 154,311 recommendation links on 40 different topics, each belonging to one of four main categories: news, science, conspiracy, and non-controversial. On the macro level, we find that conspiratorial video networks are small, sparse and centralized, with their density partly explained by lower view counts and network size. The sentiment of their videos is negative on average–like news networks, but unlike science and non-controversial networks. However, sentiment does not but view count does predict how often videos are recommended.
- 8 February 2026
- Published by Springer Nature
- p. 83-104
Open AccessThe Anatomy of Rabbit Holes: Studying Information Segregation in YouTube’s Recommendation Graph
The impact of YouTube’s recommendation algorithm has been subject to much debate. Case studies and anecdotal evidence have suggested that algorithmic recommendations may lure the platform’s users into informational ‘rabbit holes’ where they are exposed to biased content and misinformation. However, empirical research on the phenomenon has been challenging to conduct because the effects of the algorithm are inherently confounded by user and creator behavior. In this chapter, we study the net effect of algorithmic recommendations through a comparative network approach. That is, we collect the recommendation links between videos for various political and non-political issues. The resulting networks are then compared on various structural network characteristics. If the algorithm creates rabbit holes, one would expect the recommendation networks of videos on conspiratorial content to be more modular, less centralized and intra-topically dense. We analyzed 15,455 videos and 154,311 recommendation links on 40 different topics, each belonging to one of four main categories: news, science, conspiracy, and non-controversial. On the macro level, we find that conspiratorial video networks are small, sparse and centralized, with their density partly explained by lower view counts and network size. The sentiment of their videos is negative on average–like news networks, but unlike science and non-controversial networks. However, sentiment does not but view count does predict how often videos are recommended.
- Open Access
- 17 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 47-50
The Contested Commons: Governing Information-Related Public Goods Under Attack
- 17 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 47-50
Open AccessThe Contested Commons: Governing Information-Related Public Goods Under Attack
- Open Access
- 17 April 2024
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 8 (CSCW1) , 1-26
Governance Capture in a Self-Governing Community: A Qualitative Comparison of the Croatian, Serbian, Bosnian, and Serbo-Croatian Wikipedias
What types of governance arrangements make some self-governed online groups more vulnerable to disinformation campaigns? We present a qualitative comparative analysis of the Croatian and Serbian Wikipedia editions to answer this question. We do so because between at least 2011 and 2020, the Croatian language version of Wikipedia was taken over by a small group of administrators who introduced far-right bias and outright disinformation. Dissenting editorial voices were reverted, banned, and blocked. Although Serbian, Bosnian, and Serbo-Croatian Wikipedias share many linguistic and cultural features, and faced similar threats, they seem to have largely avoided this fate. Based on a grounded theory analysis of interviews with members of these communities and others in cross-functional platform-level roles, we propose that the convergence of three features---high perceived value as a target, limited early bureaucratic openness, and a preference for personalistic, informal forms of organization over formal ones---produced a window of opportunity for governance capture on Croatian Wikipedia. Our findings illustrate that online community governing infrastructures can play a crucial role in systematic disinformation campaigns and other influence operations.
- 17 April 2024
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 8 (CSCW1) , 1-26
Open AccessGovernance Capture in a Self-Governing Community: A Qualitative Comparison of the Croatian, Serbian, Bosnian, and Serbo-Croatian Wikipedias
What types of governance arrangements make some self-governed online groups more vulnerable to disinformation campaigns? We present a qualitative comparative analysis of the Croatian and Serbian Wikipedia editions to answer this question. We do so because between at least 2011 and 2020, the Croatian language version of Wikipedia was taken over by a small group of administrators who introduced far-right bias and outright disinformation. Dissenting editorial voices were reverted, banned, and blocked. Although Serbian, Bosnian, and Serbo-Croatian Wikipedias share many linguistic and cultural features, and faced similar threats, they seem to have largely avoided this fate. Based on a grounded theory analysis of interviews with members of these communities and others in cross-functional platform-level roles, we propose that the convergence of three features---high perceived value as a target, limited early bureaucratic openness, and a preference for personalistic, informal forms of organization over formal ones---produced a window of opportunity for governance capture on Croatian Wikipedia. Our findings illustrate that online community governing infrastructures can play a crucial role in systematic disinformation campaigns and other influence operations.
- Open Access
- 14 April 2023
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 7 (CSCW1) , 1-39
Mobilizing Manufactured Reality: How Participatory Disinformation Shaped Deep Stories to Catalyze Action during the 2020 U.S. Presidential Election
Claims of election fraud throughout the 2020 U.S. Presidential Election and during the lead up to the January 6, 2021 insurrection attempt have drawn attention to the urgent need to better understand how people interpret and act on disinformation. In this work, we present three primary contributions: (1) a framework for understanding the interaction between participatory disinformation and informal and tactical mobilization; (2) three case studies from the 2020 U.S. election analyzed using detailed temporal, content, and thematic analysis; and (3) a qualitative coding scheme for understanding how digital disinformation functions to mobilize online audiences. We combine resource mobilization theory with previous work examining participatory disinformation campaigns and "deep stories" to show how false or misleading information functioned to mobilize online audiences before, during, and after election day. Our analysis highlights how users on Twitter collaboratively construct and amplify alleged evidence of fraud that is used to facilitate action, both online and off. We find that mobilization is dependent on the selective amplification of false or misleading tweets by influencers, the framing around those claims, as well as the perceived credibility of their source. These processes are a self-reinforcing cycle where audiences collaborate in the construction of a misleading version of reality, which in turn leads to offline actions that are used to further reinforce a manufactured reality. Through this work, we hope to better inform future interventions.
- 14 April 2023
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Human-Computer Interaction
- Vol. 7 (CSCW1) , 1-39
Open AccessMobilizing Manufactured Reality: How Participatory Disinformation Shaped Deep Stories to Catalyze Action during the 2020 U.S. Presidential Election
Claims of election fraud throughout the 2020 U.S. Presidential Election and during the lead up to the January 6, 2021 insurrection attempt have drawn attention to the urgent need to better understand how people interpret and act on disinformation. In this work, we present three primary contributions: (1) a framework for understanding the interaction between participatory disinformation and informal and tactical mobilization; (2) three case studies from the 2020 U.S. election analyzed using detailed temporal, content, and thematic analysis; and (3) a qualitative coding scheme for understanding how digital disinformation functions to mobilize online audiences. We combine resource mobilization theory with previous work examining participatory disinformation campaigns and "deep stories" to show how false or misleading information functioned to mobilize online audiences before, during, and after election day. Our analysis highlights how users on Twitter collaboratively construct and amplify alleged evidence of fraud that is used to facilitate action, both online and off. We find that mobilization is dependent on the selective amplification of false or misleading tweets by influencers, the framing around those claims, as well as the perceived credibility of their source. These processes are a self-reinforcing cycle where audiences collaborate in the construction of a misleading version of reality, which in turn leads to offline actions that are used to further reinforce a manufactured reality. Through this work, we hope to better inform future interventions.