Early-stage concept envisioning is a critical juncture in AI design, shaping how designers frame problems and the decisions that follow. Yet values and potential harms are often too abstract or addressed too late to meaningfully shape design. Using a Research-through-Design (RtD) approach, we developed the AI Concept Envisioning Toolkit, comprising an AI Capability Library, 24 Value–Harm Cards, and a Value–Tension Map, to support reasoning by juxtaposing values and harms within AI technical capabilities. Through a survey with 30 designers and in-depth interviews with 12 designers, we find that the toolkit is clear and perceived as valuable, and that it encourages value reflection, helps anticipate potential harms, and makes ethical considerations more transparent in early-stage design. We reflect on our design process and discuss design approaches for tools that promote reflection on values and potential harms, surface and navigate value tensions, and introduce productive friction throughout design workflows.

As AI integrates into design practice, designers increasingly use generative AI tools to envision AI-enabled solutions, positioning AI as both design tool and design material. This dual role creates recursive value tensions distinct from traditional design work. We engaged 18 designers in a concept envisioning activity and interviews to understand how they navigate values and recognize potential harms in this context. Our analysis reveals that (i) designers engage in reciprocal reflection-in-action with AI; (ii) this process surfaces multi-level value tensions across tool, designer, and concept; (iii) designers demonstrate greater attunement to harm recognition as a primary design signal than to articulating positive value fulfillment; and (iv) designers exercise anticipatory judgment through meta-design reasoning about how tool assumptions risk propagating into designed concepts and future use contexts. We extend Schön’s reflection-in-action framework and discuss implications for redesigning AI-mediated design tools, supporting harm-centered reasoning, and positioning design as foundational to AI development.

Meta’s brief release of its social AI chatbots highlighted the challenges of studying systems that are both short-lived and relationally complex. In response, we conducted a 10-day collaborative autoethnography to rapidly and meaningfully engage with the product. As the first application of this method to social AI chatbots, our study demonstrates its value for examining ephemeral and emotionally complex AI systems.

In Wikipedia, conflict plays a vital role in refining knowledge. Yet some disputes persist without clear resolution or rule violations, resulting in long, circular discussions without moderator intervention. These deter neutral and third-party editors and often end only when one side gives up. In such cases, what is needed is not moderation but mediation-support that helps participants voice perspectives and reach consensus. However, mediation is hard to implement manually, as it requires summarizing complex discussions with a full grasp of both content and the Wikipedia community's culture. One potential solution is delegating this task to AI-but what content should be included in summaries that mediate and facilitate disagreements? Through a three-phase interview study with 14 Wikipedians, we examined how they read and interpret discussions, create their own summaries, and evaluate large language model (LLM)-generated summaries presented as technology probes. Our findings show that Wikipedians expect summaries to include key discussion elements-usernames, topics, arguments, sources, editor behavior, rule violations, and resolution-with particular emphasis on community-related context that reflects insider understanding. Based on these insights, we apply a new theoretical framework grounded in computer-mediated communication (CMC) theories such as common ground theory, warranting theory, and social presence theory to examine Wikipedians' interpretations of user identity. We also discuss design implications for discussion summaries on Wikipedia talk pages.

The period of transition for foster youth into independent adulthood is an important life stage, and one that has yet to be explored in HCI circles. We studied an online community centered on the experiences of former foster youth through the first year of its existence to better understand how online spaces are being used by this population. Our mixed-method study included the coding of all posts from the first year of the online community and offers a mix of quantitative and qualitative findings. These findings include alignments and gaps in an established descriptive framework from the field of social work as it relates to the online communication of former foster youth. It also includes how the domains from the framework co-occur, and some potential implications of these co-occurrences. Future research on this subject is warranted, particularly related to why former foster youth seek online platforms to engage in conversations on these topics and how effective community members perceive the platform to be in safely and securely facilitating their needs.

Do models of collaboration among contributors of Wikipedia generalize beyond the larger, western editions of the encyclopedia? In this study, we expanded upon the known collaborative mechanisms on the English Wikipedia and demonstrated that the collaboration model is best captured through the interplay of these mechanisms. We annotated talk page conversations for types of power plays or vies for control over edits that are made to articles, to understand how policy and power play mechanisms in editors' discussions account for behavior in English (EN), Farsi (FA), and Chinese (ZH) language editions of Wikipedia. Our findings show that the same power plays used in EN exist in both FA and ZH but the frequency of their usage differs across the editions. These variations suggest that editors in different language communities value contrasting types of policies to compete for power while discussing and editing articles. Our study contributes to a deeper understanding of how collaboration models developed from a western perspective translate to non-western languages.

Qualitative coding schemas are an essential part of qualitative research used in methods like Grounded Theory. To date, there is no platform to share these coding schemas. Sharing and exchanging these coding schemas has a great potential when it comes to the traceability of qualitative research and well as the re-use of coding schemas. Based on an interview study with qualitative researchers, we propose concepts for integrating a new platform for sharing qualitative coding schemas. Based on theoretical work by Birnholtz and Bietz (2002), it became clear that an easy-to-use system can foster the acceptance and the willingness of researchers to share their coding schemas. We identified three major points to focus for this on: the governance of the platform, the development of the ontology itself and integrating the sharing of qualitative coding schemas into the workflow of researchers by enabling direct upload from the qualitative coding software.

The gender gap in participation on Wikipedia is well established. The impacts of the participation bias on content may be expressed through a number of different behaviors. This research considers potential content biases that may result from efforts to delete content. We collect deletion data on a sample of article pages selected based on a method designed to identify content of likely interest to a specific group of people; men and women in this case. The analysis illustrates that there does not appear to be a systematic bias resulting in more deletions or nominations for deletions against content of likely interest to women. We consider how these results improve our understanding of bias and the ways peer production systems can mitigate the potential for biases.

Knowledge building is a prevalent feature in open online systems, but it is challenging to motivate participants to contribute and to maintain quality in the participants' contributions. Open online competitions, where participants compete for prizes with knowledge artifacts, offer a potential design model for online systems to incentivize community knowledge building activities. However, while there is evidence that participants contribute to public knowledge and share it during competitions, it remains unclear how and why they do so. In this study, through interviews with 14 participants in Kaggle Competitions, we investigate participants' motivation, practices, and challenges when contributing to community knowledge under a competitive structure. We find that competitive mechanisms impact expert and beginner participants very differently in their public knowledge building behaviors. Experts contribute to shared knowledge in order to compete for reputation, while they tend to form their own niches and only share knowledge artifacts that are abstract and not usable by less experienced participants. Beginners are often driven away from contributing to shared knowledge because of their vulnerable social image. We leverage Scardamalia's framework for Knowledge Building Communities to discuss the different challenges and opportunities that competitive design brings to expert and beginner participants. We offer design implications for effectively implementing competitive mechanisms that could benefit both expert and beginner participants in future knowledge building systems.

Although Collaborative Information-Seeking (CIS) is becoming prevalent as people engage in shared decision-making, interface components adopted in the most commonly used information seeking tools (e.g., search, filter, select, and sort) are designed for individual use. To deepen our understanding of (1) how such single-user designs affect people's consensus building processes in CIS and (2) how to devise an alternative design to improve current practices, we conducted two 4-week diary studies and observed how groups seek out places together. Our studies focus on social event coordination as a case where CIS is necessary and important. In Study 1, we examined the major challenges people encounter when performing CIS using their preferred tools. These challenges include difficulties in capturing mutual preferences, high communication cost, and disparity of work depending on a group member's perceived role as an organizer or invitee. We discovered that improving a group's shared understanding of the target information they seek (e.g., places, products) could potentially address the challenges. In Study 2, we designed, deployed, and evaluated ComeTogether, a novel system that supports a group's social event coordination. ComeTogether adopts Collaborative Dynamic Queries (C-DQ), an interface designed to allow a group to share their preferences regarding potential destinations. Study 2 results indicate that using C-DQ increased users' awareness of other group members' preferences in performing CIS, making their coordination more transparent, more inviting, and fairer than what their current practice allows. Meanwhile, ComeTogether improved communication efficiency of groups while presenting opportunities to learn about others and to discover new places. We provide implications for design that explain considerations for adopting C-DQ and identify future research directions.

Communication is critical in small group decision-making processes during which each member must be able to express preferences to reach consensus. Finding consensus can be difficult when each member in a group has a perspective that potentially conflicts with those of others. To support groups attempting to harmonize diverse preferences, we propose Collaborative Dynamic Queries (C-DQ), a UI component that enables a group to filter queries over decision criteria while being aware of others' preferences. To understand how C-DQ affects a group's behavior and perception in the decision-making process, we conducted 2 studies with groups who were prompted to make decisions together on mobile devices in a dispersed and synchronous situation. In Study 1, we found showing group preferences with C-DQ helped groups to communicate more efficiently and effectively. In Study 2, we found filtering candidates based on each member's own filter range further improved a groups' communication efficiency and effectiveness.

The English language Wikipedia is notable for its large number of articles and for the intricate collaborative interactions that create and sustain it. However, 288 other active language editions of Wikipedia have also developed through the coordination of contributing editors. While collaboration in the English Wikipedia has been researched extensively, these other language editions remain understudied. Our study leverages an influential collaboration model based on behaviors in the English Wikipedia as a lens to consider collaborative activity in the Spanish and French language editions. Through an analysis of collaborative interactions across article talk pages, we demonstrate that talk pages, the locus of most collaboration on the English Wikipedia, are used differently in these different language editions. Our study raises broader questions about how results from studies of the English Wikipedia generalize to other language editions, demonstrates the need to account for variations in collaborative behaviors in all language editions of Wikipedia and presents evidence that collaborative practices on the English Wikipedia have changed overtime.

This methods paper details an approach for identifying the representativeness of content in a user generated content (UGC) system while also accounting for endogeneity bias. We leverage metadata from an independent content provider to generate sets of commercially viable terms presumed to be of interest to specific audiences linking those terms to UGC. We describe our method and heuristics at a level of detail allowing others to follow or modify it to study both content representativeness and content gaps in UGC systems. We illustrate the method by investigating how well the English language Wikipedia addresses the content interests of four sample audiences: readers of men's and women's periodicals, and readers of political periodicals geared toward either liberal or conservative ideologies. We also share preliminary findings from each case study to demonstrate our method.

Social computing researchers are beginning to apply machine learning tools to classify and analyze social media data. Our interest in understanding politeness in an online community focused our attention on tools that would help automate politeness analysis. This paper highlights one popular classification tool designed to score the politeness of text. Our application of this tool to Wikipedia data yielded some unexpected results. Those unexpected results led us to question how the tool worked and its effectiveness relative to human judgment and classification. We designed a user study to revalidate the tool with crowdworkers labeling samples of content from Wikipedia talk pages, imitating the original efforts to validate the tool. This revalidation points to challenges for automatic labeling. Based on our results, this paper reconsiders politeness in online communities as well as broader trends in the use of machine classifiers in social computing research.

Support for coordination in online spaces, specifically in peer production systems, has frequently been an after-thought. In the absence of such support, the users of such systems must work to find an emergent order that drives shared project goals and leads to equitable processes. In short, they must rely on the "wisdom of the crowds." As our study demonstrates, however, the reality is that often the system tools available for coordination, evaluation, and work articulation are not suitable to the task at hand. Our study, first, takes a theoretical approach to understanding how tool-mediated coordination functions within peer production systems. Secondly, we enumerate the methods available to identify automated and semi-automated tools that function within such systems by quantitatively and qualitatively analyzing trace interactions and their utility in Wikipedia over a year-long period. Finally, we identify potential vacuums where new design interventions have the greatest potential for enhancing peer-production systems.

We report a study of Wikipedia in which we use a mixed-methods approach to understand how participation in specialized workgroups called WikiProjects has changed over the life of the encyclopedia. While previous work has analyzed the work of WikiProjects in supporting the development of articles within particular subject domains, the collaborative role of WikiProjects that do not fit this conventional mold has not been empirically examined. We combine content analysis, interviews and analysis of edit logs to identify and characterize these alternative WikiProjects and the work they do. Our findings suggest that WikiProject participation reflects community concerns and shifts in the community's conception of valued work over the past six years. We discuss implications for other open collaborations that need flexible, adaptable coordination mechanisms to support a range of content creation, curation and community maintenance tasks.

In online groups, increasing explicit coordination can increase group cohesion and member productivity. On Wikipedia, groups called WikiProjects employ a variety of explicit coordination mechanisms to motivate and structure member contribution, with the goal of creating and improving articles related to particular topics. However, while explicit coordination works well for coordinating article-level actions, coordinating group tasks and tracking progress towards group goals that involve tracking hundreds or thousands of articles over time requires different coordination strategies.

WikiProjects have contributed to Wikipedia's success in important ways, yet the range of work that WikiProjects perform and the way they coordinate that work remains largely unexplored. In this study, we perform a content analysis of 788 work-related discussions from the talk pages of 138 WikiProjects in order to understand the role WikiProjects play in collaborative work on Wikipedia. We find that the editors use WikiProjects to coordinate a wide variety of work activities beyond content production and that non-members play an active role in that work. Our research suggests that WikiProject collaboration is less structured and more open than that of many virtual teams and that WikiProjects may function more like FLOSS projects than traditional groups.

The relationships and work that facilitate content creation in large online contributor system are not always visible. Social translucence is a stance toward the design of systems that allows users to better understand collaborative system participation through awareness of contributions and interactions. Like many socio-technical constructs, social translucence is not something that can be simply added after a system is built; it should be at the core of system design. In this paper, we conduct a domain analysis to understand the space of architectural support required to facilitate social translucence in systems. We describe an instantiation of those requirements as a system architecture that relies on data from Wikipedia and illustrate how translucence can be propagated to some basic visualizations which we have created for Wikipedia users. We close with some reflections on the state of social translucence research and some openings for this important design perspective.

This poster introduces Qbox, a flexible tool developed by the Communicative Practices in Virtual Workspaces research group at the University of Washington to support traditional and innovative forms of analysis for web-based and digital material. Qbox integrates three functional areas of work associated with content analysis: consolidating and presenting source data, performing coding or classification work, and analyzing results. Developed using an iterative user-centered design approach to support ongoing research, this tool enhances research protocol by providing a flexible application to organize digital spaces, and demonstrates the power of productivity associated with agile, user-based development.

Groups collaborating in online spaces on complex, extended projects develop behavioral conventions and agreed-upon practices to structure and regulate their interactions and work. Collaborators on Wikipedia have developed a multi-tiered policy environment to document a set of evolving principles, processes, and rules to facilitate productive group collaboration. Previous quantitative studies have noted this hierarchical structure, but have evaluated the policy environment as a singular entity rather than investigating potential differences between the three main regulatory genres that enable it. These studies also excluded essays, the least official regulatory genre, from their analyses. We perform a comparative content analysis of all three genres (policies, guidelines, and essays) and demonstrate that they focus on different areas of community regulation. Drawing on the theory of genre ecologies we discuss the possible role of unofficial genres such as essays in articulating and regulating work practices in online, organized collaborative work.

This paper looks at the problem of detecting a particular type of social behavior in discussions: attempts to establish credibility as an authority on a particular topic. Using maximum entropy modeling, we explore questions related to feature extraction and turn vs. discussion-level modeling in experiments with online discussion text given only a small amount of labeled training data. We also introduce a method for learning interaction words from unlabeled data. Preliminary experiments show that a word-based approach (as used in topic classification) can be used successfully for turn-level modeling, but is less effective at the discussion level. We also find that sentence complexity features are almost as useful as lexical features, and that interaction words are more robust than the full vocabulary when combined with other features.