Search

As AI systems quickly improve in both breadth and depth of performance, they lend themselves to creating increasingly powerful and realistic agents, including the possibility of agents modeled on specific people. We anticipate that within our lifetimes it may become common practice for people to create custom AI agents to interact with loved ones and/or the broader world after death; indeed, the past year has seen a boom in startups purporting to offer such services. We call these generative ghosts, since such agents will be capable of generating novel content rather than merely parroting content produced by their creator while living. In this paper, we reflect on the history of technologies for AI afterlives, including current early attempts by individual enthusiasts and startup companies to create generative ghosts. We then introduce a novel design space detailing potential implementations of generative ghosts, and use this analytic framework to ground discussion of the practical and ethical implications of various approaches to designing generative ghosts, including potential positive and negative impacts on individuals and society. Based on these considerations, we lay out a research agenda for the AI and HCI research communities to better understand the risk/benefit landscape of this novel technology so as to ultimately empower people who wish to create and interact with AI afterlives to do so in a beneficial manner., Comment: version 4, updated to include new references and examples

Interpretive scholars generate knowledge from text corpora by manually sampling documents, applying codes, and refining and collating codes into categories until meaningful themes emerge. Given a large corpus, machine learning could help scale this data sampling and analysis, but prior research shows that experts are generally concerned about algorithms potentially disrupting or driving interpretive scholarship. We take a human-centered design approach to addressing concerns around machine-assisted interpretive research to build Scholastic, which incorporates a machine-in-the-loop clustering algorithm to scaffold interpretive text analysis. As a scholar applies codes to documents and refines them, the resulting coding schema serves as structured metadata which constrains hierarchical document and word clusters inferred from the corpus. Interactive visualizations of these clusters can help scholars strategically sample documents further toward insights. Scholastic demonstrates how human-centered algorithm design and visualizations employing familiar metaphors can support inductive and interpretive research methodologies through interactive topic modeling and document clustering., Comment: To appear at the 2022 ACM Symposium on User Interface Software and Technology (UIST '22)

Content moderation research typically prioritizes representing and addressing challenges for one group of stakeholders or communities in one type of context. While taking a focused approach is reasonable or even favorable for empirical case studies, it does not address how content moderation works in multiple contexts. Through a systematic literature review of 86 content moderation papers that document empirical studies, we seek to uncover patterns and tensions within past content moderation research. We find that content moderation can be characterized as a series of trade-offs around moderation actions, styles, philosophies, and values. We discuss how facilitating cooperation and preventing abuse, two key elements in Grimmelmann's definition of moderation, are inherently dialectical in practice. We close by showing how researchers, designers, and moderators can use our framework of trade-offs in their own work, and arguing that trade-offs should be of central importance in investigating and designing content moderation., Comment: To appear in ACM TOCHI

The proliferation of harmful content on online social media platforms has necessitated empirical understandings of experiences of harm online and the development of practices for harm mitigation. Both understandings of harm and approaches to mitigating that harm, often through content moderation, have implicitly embedded frameworks of prioritization - what forms of harm should be researched, how policy on harmful content should be implemented, and how harmful content should be moderated. To aid efforts of better understanding the variety of online harms, how they relate to one another, and how to prioritize harms relevant to research, policy, and practice, we present a theoretical framework of severity for harmful online content. By employing a grounded theory approach, we developed a framework of severity based on interviews and card-sorting activities conducted with 52 participants over the course of ten months. Through our analysis, we identified four Types of Harm (physical, emotional, relational, and financial) and eight Dimensions along which the severity of harm can be understood (perspectives, intent, agency, experience, scale, urgency, vulnerability, sphere). We describe how our framework can be applied to both research and policy settings towards deeper understandings of specific forms of harm (e.g., harassment) and prioritization frameworks when implementing policies encompassing many forms of harm., Comment: CSCW 2021; 33 pages

Qualitative inductive methods are widely used in CSCW and HCI research for their ability to generatively discover deep and contextualized insights, but these inherently manual and human-resource-intensive processes are often infeasible for analyzing large corpora. Researchers have been increasingly interested in ways to apply qualitative methods to "big" data problems, hoping to achieve more generalizable results from larger amounts of data while preserving the depth and richness of qualitative methods. In this paper, we describe a study of qualitative researchers' work practices and their challenges, with an eye towards whether this is an appropriate domain for human-AI collaboration and what successful collaborations might entail. Our findings characterize participants' diverse methodological practices and nuanced collaboration dynamics, and identify areas where they might benefit from AI-based tools. While participants highlight the messiness and uncertainty of qualitative inductive analysis, they still want full agency over the process and believe that AI should not interfere. Our study provides a deep investigation of task delegability in human-AI collaboration in the context of qualitative analysis, and offers directions for the design of AI assistance that honor serendipity, human agency, and ambiguity., Comment: 23 pages. Accepted to ACM CSCW 2021

Online community moderators are on the front lines of combating problems like hate speech and harassment, but new modes of interaction can introduce unexpected challenges. In this paper, we consider moderation practices and challenges in the context of real-time, voice-based communication through 25 in-depth interviews with moderators on Discord. Our findings suggest that the affordances of voice-based online communities change what it means to moderate content and interactions. Not only are there new ways to break rules that moderators of text-based communities find unfamiliar, such as disruptive noise and voice raiding, but acquiring evidence of rule-breaking behaviors is also more difficult due to the ephemerality of real-time voice. While moderators have developed new moderation strategies, these strategies are limited and often based on hearsay and first impressions, resulting in problems ranging from unsuccessful moderation to false accusations. Based on these findings, we discuss how voice communication complicates current understandings and assumptions about moderation, and outline ways that platform designers and administrators can design technology to facilitate moderation., Comment: 23 pages

User-generated content is central to social computing scholarship. However, researchers and practitioners often presume that these users are alive. Failing to account for mortality is problematic in social media where an increasing number of profiles represent those who have died. Identifying mortality can empower designers to better manage content and support the bereaved, as well as promote high-quality data science. Based on a computational linguistic analysis of post-mortem social media profiles and content, we report on classifiers developed to detect mortality and show that mortality can be determined after the first few occurrences of post-mortem content. Applying our classifiers to content from two other platforms also provided good results. Finally, we discuss trade-offs between models that emphasize pre- vs. post-mortem precision in this sensitive context. These results mark a first step toward identifying mortality at scale, and show how designers and scientists can attend to mortality in their work.

Animated GIFs are increasingly popular in text-based communication. Finding the perfect GIF can make conversations funny, interesting, and engaging, but GIFs also introduce potentials for miscommunication. Through 24 in-depth qualitative interviews, this empirical, exploratory study examines the nuances of communication practices with animated GIFs to better understand why and how GIFs can send unintentional messages. We find participants leverage contexts like source material and interpersonal relationship to find the perfect GIFs for different communication scenarios, while these contexts are also the primary reason for miscommunication and some technical usability issues in GIFs. This paper concludes with a discussion of the important role that different types of context play in the use and interpretations of GIFs, and argues that nonverbal communication tools should account for complex contexts and common ground that communication media rely on.