Controlling Polarization in Personalization

Celis, L. Elisa; Kapoor, Sayash; Salehi, Farnood; Vishnoi, Nisheeth K.

doi:10.1145/3287560.3287601

Cited by 68 publications

(56 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…But this reinforcement bias is not the only problem of today's algorithm-controlled dissemination of information. The increased levels of personalization can also lead to filter bubbles (Pariser 2011) and echo chambers (Celis et al 2019). The reason is that modern algorithms learn over time which type of content a certain user likes or dislikes.…”

Section: Selected Disciplinary Perspectives On Infodemics and The Potmentioning

confidence: 99%

Understanding the dynamics emerging from infodemics: a call to action for interdisciplinary research

et al. 2021

View full text Add to dashboard Cite

Research on infodemics, i.e., the rapid spread of (mis)information related to a hazardous event, such as the COVID-19 pandemic, requires integrating a multiplicity of scientific disciplines. The dynamics emerging from infodemics have the potential to generate complex behavioral patterns. To react appropriately, it is of ultimate importance for the fields of Business and Economics to understand these dynamics. In the short run, they might lead to an adaptation in household spending or to a shift in buying behavior towards online providers. In the long run, changes in investments, consumer behavior, and markets are to be expected. We argue that the dynamics emerge from complex interactions among multiple factors, such as information and misinformation accessible to individuals and the formation and revision of beliefs. (Mis)information accessible to individuals is, amongst others, affected by algorithms specifically designed to provide personalized information, while automated fact-checking algorithms can help reduce the amount of circulating misinformation. The formation and revision of individual (and probably false) beliefs and individual fact-checking and interpretation of information are heavily affected by linguistic patterns inherent to information during pandemics and infodemics and further factors, such as affect, intuition, and motives. We argue that, to get a deep(er) understanding of the dynamics emerging from infodemics, the fields of Business and Economics should integrate the perspectives of Computer Science and Information Systems, (Computational) Linguistics, and Cognitive Science into the wider context of economic systems (e.g., organizations, markets or industries) and propose a way to do so. As research on infodemics is a strongly interdisciplinary field and the integration of the above-mentioned disciplines is a first step towards a holistic approach, we conclude with a call to action which should encourage researchers to collaborate across scientific disciplines and unfold collective creativity, which will substantially advance research on infodemics.

show abstract

Section: Selected Disciplinary Perspectives On Infodemics and The Potmentioning

confidence: 99%

Understanding the dynamics emerging from infodemics: a call to action for interdisciplinary research

et al. 2021

View full text Add to dashboard Cite

show abstract

“…These works focus on diversity of content consumption, showing that it can be increased by natural collaborative filtering based recommendations. Other works propose remedies to 'fix' recommender systems to limit their polarizing effect [CKSV19,HMRU21]. It would be interesting to understand the effects of these remedies in our opinion and edge dynamics model.…”

Section: Related Workmentioning

confidence: 99%

Local Edge Dynamics and Opinion Polarization

Bhalla¹,

Lechowicz²,

Musco³

2021

Preprint

View full text Add to dashboard Cite

The proliferation of social media platforms, recommender systems, and their joint societal impacts have prompted significant interest in opinion formation and evolution within social networks. In this work, we study how local dynamics in a network can drive opinion polarization. In particular, we study time evolving networks under the classic Friedkin-Johnsen opinion model. Edges are iteratively added or deleted according to simple local rules, modeling decisions based on individual preferences and network recommendations. 1 We give theoretical bounds showing how individual edge updates affect polarization, and a related measure of disagreement across edges. Via simulations on synthetic and real-world graphs, we find that the presence of two simple dynamics gives rise to high polarization: 1) confirmation bias -i.e., the preference for nodes to connect to other nodes with similar expressed opinions and 2) friend-of-friend link recommendations, which encourage new connections between closely connected nodes. We also investigate the role of fixed connections which are not subject to these dynamics. We find that even a small number of fixed edges can significantly limit polarization, but still lead to multimodal opinion distributions, which may be considered polarized in a different sense.

show abstract

“…Social media companies do moderate their content to screen out child pornography, extreme violence, etc., although even this is quite difficult. Yet, emotionally charged subject matter drives engagement (Celis et al 2019). This dovetails with the need to attract consumers to the site so that they can be marketed to advertisers.…”

Section: Lack Of Investment In Moderationmentioning

confidence: 99%

The easy weaponization of social media: why profit has trumped security for U.S. companies

Oates

2020

Digi War

View full text Add to dashboard Cite

American-based social media companies have become active players in digital war, both by accident of design and a subsequent failure to address the threat due to concerns over profits. Discussions about the negative role of social media in society generally address the myriad problems wrought by social media, including electoral manipulation, foreign disinformation, trolling, and deepfakes, as unfortunate side effects of a democratizing technology. This article argues that the design of social media fosters information warfare. With its current composition and lack of regulation, social media platforms such as Facebook and Twitter are active agents of disinformation, their destructive force in society outweighing their contributions to democracy. While this is not by deliberate design, the twin forces of capitalism and a lack of regulation of the world's largest social media platforms have led to a situation in which social media are a key component of information war around the globe. This means that scholarly discussions should shift away from questions of ethics or actions (or lack thereof) on the part of social media companies to a frank focus on the security risk posed to democracy by social media.

show abstract

Controlling Polarization in Personalization

Cited by 68 publications

References 22 publications

Understanding the dynamics emerging from infodemics: a call to action for interdisciplinary research

Understanding the dynamics emerging from infodemics: a call to action for interdisciplinary research

Local Edge Dynamics and Opinion Polarization

The easy weaponization of social media: why profit has trumped security for U.S. companies

Contact Info

Product

Resources

About