Bridging Systems: Open Problems for Countering Destructive Divisiveness across Ranking, Recommenders, and Governance

Abstract

Divisiveness appears to be increasing in much of the world, leading to concern about political violence and a decreasing capacity to collaboratively address large-scale societal challenges. In this working paper we aim to articulate an interdisciplinary research and practice area focused around what we call bridging systems: systems which increase mutual understanding and trust across divides, creating space for productive conflict, deliberation, or cooperation. We give examples of bridging systems across three domains: recommender systems on social media, software for conducting civic forums, and human-facilitated group deliberation. We argue that these examples can be more meaningfully understood as processes for attention-allocation (as opposed to "content distribution" or "amplification"), and develop a corresponding framework to explore similarities-and opportunities for bridging-across these seemingly disparate domains. We focus particularly on the potential of bridging-based ranking to bring the benefits of offline bridging into spaces which are already governed by algorithms. Throughout, we suggest research directions that could improve our capacity to incorporate bridging into a world increasingly mediated by algorithms and artificial intelligence.

The Evolution of Online News Headlines

“…Mesoudi & Thornton, 2018, for cumulative cultural evolution), the arbitrariness of online success metrics presents an opportunity to design sustainable, meaningful online metrics that benefit society and the individual: e.g., the diversity of readership may indicate higher quality (Bhadani et al, 2022), or algorithms may act as bridging systems (Ovadya & Thorburn, 2023). The Guardian now displays a "Deeply read" list next to its "Most viewed" list on its homepage, offering readers an additional criterion for deciding where to spend their time online ('What Is the "Deeply Read" List?…”

Mapping the Design Space of Teachable Social Media Feed Experiences

“…Here, we focus on a particular class of content algorithms ubiquitous on social media and other everyday applications-recommender systems. A growing body of work proposes techniques and broader calls to action to embed human values into recommender systems [21,36,64,82,103,104], drawing upon the field of value-sensitive design [11,35]. Borning and Muller define "value" as "what a person or group of people consider important in life" [11].…”

The Algorithmic Management of Polarization and Violence on Social Media

“…Bridging-based ranking prioritizes content that meets approval (or generates positive engagement) across diverse groups of people. This approach attempts to counteract the amplification of divisive material by favoring items which have cross-partisan appeal (Ovadya and Thorburn 2023). A simple example is Facebook's use of a crowdsourced survey to rate the credibility of news domains, rating as trustworthy only those with a supermajority of support (Owen, 2018).…”