To date, efforts to address misinformation have focused on broadly scoped legal, technical, and educational interventions. However, recent studies in HCI have begun exploring new directions that focus instead on the individualized nature of the problem and its roots in identity and community. To understand the potential benefits and challenges of applying community-driven methods in this and similarly contentious spaces, we examined the work of a large-scale participatory research project working to address the impact of misinformation, predominantly on rural communities and BIPOC (Black, Indigenous, and People of Color) individuals, through the development of locally-contextualized educational resources and platforms. Through interviews with 23 project stakeholders, including both community partners and researchers, we heard that participants saw community-based research and design methods as a highly beneficial tool for addressing this problem. They shared that such methods allow participants to build relationships through the sharing of complex emotions and to surface a range of expertise and lived experiences that contribute to a more holistic understanding of misinformation. At the same time, participatory, community-based methods raise difficult questions around integrating such diverse expertise and protecting participants from internal and external risk. We also discuss emergent strategies for navigating these tensions, with the aim of informing and preparing CSCW researchers looking to apply participatory methods towards addressing misinformation and other contentious research topics.

Toolkits are an important means of sharing expertise and influencing practice. However, the work of making and sustaining toolkits is not well understood. We address this gap by conducting 20 semi-structured interviews with toolkit designers, focusing on toolkits intended to help practitioners such as librarians, teachers, and community workers. We analyze these interviews to surface key aspects of participants’ design journeys: (1) how their projects began; (2) how they conceptualized use; (3) how they collaborated with users; (4) and what happened once their toolkit was released. We illustrate these aspects through three narratives, and discuss our findings to provide considerations for designers and scholars. We highlight how designers co-construct communities alongside their toolkits, helping us form a more nuanced understanding of the social aspects underpinning toolkit projects. Collectively, these contributions can help us identify challenges and opportunities in this design space, laying the groundwork to increase toolkits’ social impact.

Social computing researchers have described how online harassment is taking on more sophisticated forms where communities rather than lone individuals carry out attacks. At the same time, those who are targeted are coming together with members of the online crowd to confront this networked harassment. However, we currently lack empirical studies of how such self-organizing efforts take place, especially in non-western contexts like India. To address this gap, this paper examines two case studies of gendered Islamophobia, Sulli Deals and Bulli Bai where a group of Indian Muslim women were targeted via fake auctions hosted on GitHub, that were subsequently amplified over social media. We conducted 12 in-depth interviews with the women targeted in these incidents to understand how these women collectively made sense of these incidents, and the social and spiritual resources they drew upon as they constructed a response together. Our findings describe how our participants understood these incidents (e.g. with families and each other); the work they did to combat the harassment (e.g. coordinating a unified media response); and how religious beliefs and spiritual practices played into this (e.g. values around courage, patience, flexibility, and humility). We conclude by discussing how these findings can refresh our understanding of designing community-driven responses to networked harassment.

Recent studies in HCI have explored how we might reduce the spread of online misinformation by helping people learn how to evaluate information in more skillful ways. Unfortunately, it isn’t clear that such interventions have been meaningfully integrated into communities. To better understand why this is the case, this paper engages over thirty information professionals (educators, librarians, and journalists) who promote digital literacy in BIPOC and rural communities. Our participants describe a temporal mismatch, whereby digital literacy requires time-consuming processes that cannot be accelerated, but institutional and societal pressures demand speed. We also describe strategies that participants envisaged to cope with this mismatch. This leads us to discuss how the HCI community can better engage with the temporal aspects of digital literacy work, with a view toward expanding the range of solutions we can design to address the misinformation crisis.

In this paper, we argue that strategic information operations (e.g. disinformation, political propaganda, and other forms of online manipulation) are a critical concern for CSCW researchers, and that the CSCW community can provide vital insight into understanding how these operations function-by examining them as collaborative "work" within online crowds. First, we provide needed definitions and a framework for conceptualizing strategic information operations, highlighting related literatures and noting historical context. Next, we examine three case studies of online information operations using a sociotechnical lens that draws on CSCW theories and methods to account for the mutual shaping of technology, social structure, and human action. Through this lens, we contribute a more nuanced understanding of these operations (beyond "bots" and "trolls") and highlight a persistent challenge for researchers, platform designers, and policy makers-distinguishing between orchestrated, explicitly coordinated, information operations and the emergent, organic behaviors of an online crowd.

This research examines how Russian disinformation actors participated in a highly charged online conversation about the #BlackLivesMatter movement and police-related shootings in the USA during 2016. We first present high-level dynamics of this conversation on Twitter using a network graph based on retweet flows that reveals two structurally distinct communities. Next, we identify accounts in this graph that were suspended by Twitter for being affiliated with the Internet Research Agency, an entity accused of conducting information operations in support of Russian political interests. Finally, we conduct an interpretive analysis that consolidates observations about the activities of these accounts. Our findings have implications for platforms seeking to develop mechanisms for determining authenticity---by illuminating how disinformation actors enact authentic personas and caricatures to target different audiences. This work also sheds light on how these actors systematically manipulate politically active online communities by amplifying diverging streams of divisive content.

This paper examines how users of social media correct online rumors during crisis events. Focusing on Twitter, we identify different patterns of information correcting behaviors and describe the actions, motivations, rationalizations and experiences of people who exhibited them. To do this, we analyze digital traces across two separate crisis events and interviews of fifteen individuals who generated some of those traces. Salient themes ensuing from this work help us describe: 1) different mechanisms of corrective action with respect to who gets corrected and how; 2) how responsibility is positioned for verifying and correcting information; and 3) how users' imagined audience influences their corrective strategy. We synthesize these three components into a preliminary model and explore the role of imagined audiences-both who those audiences are and how they react to and interact with shared information-in shaping users' decisions about whether and how to correct rumors.

This research examines Twitter discourse related to #BlackLivesMatter and police-related shooting events in 2016 through a mixed-method, interpretative approach. We construct a "shared audience graph", revealing structural and ideological disparities between two groups of participants (one on the political left, the other on the political right). We utilize an integrated networked gatekeeping and framing lens to examine how #BlackLivesMatter frames were produced--and how they were contested --by separate communities of supporters and critics. Among other empirical findings, this work demonstrates hashtags being used in diverse ways--e.g. to mark participation, assert individual identity, promote group identity, and support or challenge a frame. Considered from a networked gatekeeping perspective, we illustrate how hashtags can serve as channeling mechanisms, shaping trajectories of information flow. This analysis also reveals a right-leaning community of BlackLivesMatter critics to have a more well-defined group of crowdsourced elite who largely define their side's counter-frame.

In this paper we highlight three distinct approaches to studying rumor dynamics-volume, exposure, and content production. Expanding upon prior work, which has focused on rumor volume, we argue that considering the size of the exposed population is a vital component of understanding rumoring. Additionally, by combining all three approaches we discover subtle features of rumoring behavior that would have been missed by applying each approach in isolation. Using a case study of rumoring on Twitter during a hostage crisis in Sydney, Australia, we apply a mixed-methods framework to explore rumoring and its consequences through these three lenses, focusing on the added dimension of exposure in particular. Our approach demonstrates the importance of considering both rumor content and the people engaging with rumor content to arrive at a more holistic understanding of communication dynamics. These results have implications for emergency responders and official use of social media during crisis management.