HCI research has been at the forefront of designing interventions for protecting teens online; yet, how can we test and evaluate these solutions without endangering the youth we aim to protect? Towards this goal, we conducted focus groups with 20 teens to inform the design of a social media simulation platform and study for evaluating online safety nudges co-designed with teens. Participants evaluated risk scenarios, personas, platform features, and our research design to provide insight regarding the ecological validity of these artifacts. Teens expected risk scenarios to be subtle and tricky, while also higher in risk to be believable. The teens iterated on the nudges to prioritize risk prevention without reducing autonomy, risk coping, and community accountability. For the simulation, teens recommended using transparency with some deceit to balance realism and respect for participants. Our meta-level research provides a teen-centered action plan to evaluate online safety interventions safely and effectively.

This study addresses two currently open questions about how behaviors of online community members influence others’ responses to misinformation. First, in contrast to prior work, it directly measures norm perception to address whether (1) norm perception actually acts as a mediator, (2) others’ behaviors directly influence individuals’ responses to misinformation, (3) both direct and mediated effects occur. Second, it investigates norm perceptions about a behavior that is not readily observable in online communities, but is prone to misinformation, specifically, vaccination. To do so, it experimentally manipulates the prevalence of communicating about vaccination (an unobservable behavior) within an online community. The results demonstrate no evidence of a direct effect—the causal relationship between prevalence of communicating a behavior and intentions to respond to misinformation only occurs via norm perception as a mediator. The paper highlights implications of these findings for designing community-centered interventions to influence perceived norms, thereby mitigating misinformation spread and impacts.

Social norms play a significant role in how conspiratorial content and related misinformation impact online communities. However, less is understood about the mechanisms by which particular aspects of a community may drive perceptions of social norms in the community. Using anti-vaccine conspiracies as a testbed, this paper experimentally examines three such features and their relationships : prevalence of conspiratorial content, community response, and explicit community rules. Results show that prevalence of content has a significant effect on norm perceptions, while the results did not support the effects of explicit rule on norm perceptions. However, these effects can be mitigated by the way a community responds to such content. Furthermore, perceived norms also influence other expectations about the community, from escalated behaviors to belief in other conspiracy theories. The paper concludes by highlighting the implications of these findings for online platform design, for community governance, and for future research about the relationships among conspiratorial content and norm perceptions.

Prior work has identified a variety of factors that drive the way people identify and respond to misinformation. Such factors include confirmation bias, perceived credibility of the information source, individual media literacy, social norms, and others. This paper reviews the interventions designed to address misinformation and examines how various underlying mechanisms of response to misinformation are operationalized and implemented in the reviewed interventions. Key findings show that most prior work to address misinformation heavily focuses on individual pieces of misinformation and the actions individuals take in response to those individual pieces. These individualistic approaches, we argue, overlook the other drivers of responses to misinformation, such as individuals' prior beliefs and the social contexts in which misinformation is encountered. Additionally, the analysis shows that an individualistic focus on misinformation draws attention away from the systemic nature and consequences of misinformation. This paper argues that to overcome the limitation of individualistic approaches to addressing misinformation, future interventions need to expand their scope beyond individualistic approaches. As one way to do so, it discusses leveraging the impacts of community factors that impact the spread and impacts of misinformation. The paper concludes by using social norms as an example to illustrate how a focus on community factors might work in practice.

As AI-mediated communication (AI-MC) becomes more prevalent in everyday interactions, it becomes increasingly important to develop a rigorous understanding of its effects on interpersonal relationships and on society at large. Controlled experimental studies offer a key means of developing such an understanding, but various complexities make it difficult for experimental AI-MC research to simultaneously achieve the criteria of experimental realism, experimental control, and scalability. After outlining these methodological challenges, this paper offers the concept of methodological middle spaces as a means to address these challenges. This concept suggests that the key to simultaneously achieving all three of these criteria is to abandon the perfect attainment of any single criterion. This concept's utility is demonstrated via its use to guide the design of a platform for conducting text-based AI-MC experiments. Through a series of three example studies, the paper illustrates how the concept of methodological middle spaces can inform the design of specific experimental methods. Doing so enabled these studies to examine research questions that would have been either difficult or impossible to investigate using existing approaches. The paper concludes by describing how future research could similarly apply the concept of methodological middle spaces to expand methodological possibilities for AI-MC research in ways that enable contributions not currently possible.

Research on human-robot collaboration or human-robot teaming, has focused predominantly on understanding and enabling collaboration between a single robot and a single human. Extending human-robot collaboration research beyond the dyad, raises novel questions about how a robot should allocate resources among group members and about what the consequences of such allocation are for a group’s social dynamics and outcomes. Methodological advances are needed to answer these questions allow researchers to collect data about a robot’s impact not only on interactions with the robot but also on interactions of people with each other. This paper presents Robot Assisted Tower Construction, a novel task that allows researchers to examine the impact of a robot’s allocation behavior on the dynamics of a group or team collaborating on a task. By focusing on the question of whether and how a robot’s allocation of resources (wooden blocks required for a building task) affects collaboration dynamics and outcomes, a case is provided of how this task can be applied in a laboratory study with 124 participants to collect data about human robot collaboration that involves a group of people. We highlight the kinds of insights the task can yield and how it can be adapted to various human robot collaboration contexts.

Conversational agents are increasingly becoming integrated into everyday technologies and can collect large amounts of data about users. As these agents mimic interpersonal interactions, we draw on communication privacy management theory to explore people's privacy expectations with conversational agents. We conducted a 3x3 factorial experiment in which we manipulated agents' social interactivity and data sharing practices to understand how these factors influence people's judgments about potential privacy violations and their evaluations of agents. Participants perceived agents that shared response data with advertisers more negatively compared to agents that shared such data with only their companies; perceptions of privacy violations did not differ between agents that shared data with their companies and agents that did not share information at all. Participants also perceived the socially interactive agent's sharing practices less negatively than those of the other agents, highlighting a potential privacy vulnerability that users are exposed to in interactions with socially interactive conversational agents.

Social media sites are where life happens for many of today's young people, so it is important to teach them to use these sites safely and effectively. Many youth receive classroom education on digital literacy topics, but have few chances to build actual skills. Social Media TestDrive, an interactive social media simulation, fills a gap in digital literacy education by combining experiential learning in a realistic and safe social media environment with educator-facilitated classroom lessons. The tool was piloted with 12 educators and over 200 students, and formative evaluation data suggest that TestDrive achieved high levels of engagement with both groups. Students reported the modules enhanced their understanding of digital citizenship issues, and educators noted that students were engaging in meaningful classroom conversations. Finally, we discuss the importance of involving multiple stakeholder groups (e.g., researchers, youth, educators, curriculum developers) in designing educational technology.

Bystander intervention can reduce the amount of cyberbullying victimization on social media, but bystanders often fail to act. Limited accountability for their behavior and a lack of empathy for the victim are frequently cited as reasons for why bystanders do not act against cyberbullying. We developed design interventions that aimed to increase accountability and empathy among bystanders. In Study 1, participants were experimentally exposed to three social media posts with different types of empathy nudges. Empathy nudges embedded into social media posts displayed the potential to motivate empathy. In Study 2, participants took part in a 3-day experiment that simulated a social media experience. Results suggested that increased social transparency on social media promoted accountability through heightened self-presentation concerns, but empathy nudges did not encourage greater bystander empathy. Both accountability and empathy predicted bystander intervention, but the types of bystander actions promoted by each mechanism differed. We consider how these results contribute to theories of bystander behavior and designing social media to promote prosocial behaviors.

Although bystander intervention can mitigate the negative effects of cyberbullying, few bystanders ever attempt to intervene. In this study, we explored the effects of interface design on bystander intervention using a simulated custom-made social media platform. Participants took part in a three-day, in-situ experiment, in which they were exposed to several cyberbullying incidents. Depending on the experimental condition, they received different information about the audience size and viewing notifications intended to increase a sense of personal responsibility in bystanders. Results indicated that bystanders were more likely to intervene indirectly than directly, and information about the audience size and viewership increased the likelihood of flagging cyberbullying posts through serial mediation of public surveillance, accountability, and personal responsibility. The study has implications for understanding bystander effect in cyberbullying, and how to develop design solutions to encourage bystander intervention in social media.

Virtual worlds present a natural test bed to observe and study the social behaviors of people at a large scale. Analyzing the rich “big data” generated by the activities of players in a virtual world enables us to better understand the online society, to validate and propose sociological theories, and to provide insights of how people behave in the real world. However, how to better store and analyze such complex big data has always been an issue that prevents in-depth analyses. In this paper, we first review the research on trust in virtual worlds and Semantic Web as applied in social network analysis. Then, we present how we employed Semantic Web technologies to address this issue, and how we explored certain social concepts expressed within a massively multiplayer online game-Ever Quest I I. Specifically, the relations between mentors and mentees in the game are studied. We use the housing network to measure the trust between mentors and mentees and adopt the logistic regression model to identify the predictors of building trust between them. Our research sheds light on how to analyze largescale data within a virtual world by exploring the flow of trust in different layers of social networks with the help of Semantic Web technologies.

The Open Government Directive is making US government data available via websites such as Data.gov for public access. In this paper, we present a Semantic Web based approach that incrementally generates Linked Government Data (LGD) for the US government. In focusing on the trade-off between high quality LGD generation (requiring non-trivial human expert input) and massive LGD generation (requiring low human processing cost), our work is highlighted by the following features: (i) supporting low-cost and extensible LGD publishing for massive government data; (ii) using Social Semantic Web (Web3.0) technologies to incrementally enhance published LGD via crowdsourcing, and (iii) facilitating mash-ups by declaratively reusing cross-dataset mappings which usually are hard-coded in applications.