Why do social computing projects aimed at alleviating social inequality fail? This paper investigates this question through a qualitative interview study with 25 individuals working to address the problem of wage theft in the United States (US) context. Our analyses uncover failures at three levels or scales of interaction: one, failures at the individual level of technology adoption; two, relational failures (i.e., the anti-labor worker/employer dynamic in the US); and three, institutional or macro-level failures. Taken together, these various failings point to larger, structural forces that negatively fate pro-labor projects’ trajectories – i.e., capitalism. Capitalism's incarnations in the US play a significant and at times harsh grip in steering the path of social computing design projects. In this paper, we untangle the relationship between capitalism and social computing, providing an analytic framework to tease apart this complex relationship, the lessons learned from our empirical data, as well as ways forward for future, pro-labor, social computing projects.

Play is an essential part of the human experience and can be found throughout the lifespan. While play has long been of interest to the HCI community, research has often focused on the technologies supporting game play, the potential outcomes of play (e.g., skill-building, health improvements), or play among children. This paper explores what play looks like in online communities that are not specifically game-based and consist primarily of adults. From online ethnographic work of the ARMY (i.e., Adorable Representative M.C. for Youth), fandom of the South Korean musical group BTS, we explore how BTS and ARMY collaboratively construct a playful social environment using various social media platforms. A contribution of this work is to expand our conceptualization of how adults create playful places that are not specifically game-based and highlights the role of socio-technical systems in their community building.

The interpretability or explainability of AI systems (XAI) has been a topic gaining renewed attention in recent years across AI and HCI communities. Recent work has drawn attention to the emergent explainability requirements of in situ, applied projects, yet further exploratory work is needed to more fully understand this space. This paper investigates applied AI projects and reports on a qualitative interview study of individuals working on AI projects at a large technology and consulting company. Presenting an empirical understanding of the range of stakeholders in industrial AI projects, this paper also draws out the emergent explainability practices that arise as these projects unfold, highlighting the range of explanation audiences (who), as well as how their explainability needs evolve across the AI project lifecycle (when). We discuss the importance of adopting a sociotechnical lens in designing AI systems, noting how the “AI lifecycle” can serve as a design metaphor to further the XAI design field.

Ground-truth labeling is an important activity in machine learning. Many studies have examined how crowdworkers apply labels to records in machine learning datasets. However, there have been few studies that have examined the work of domain experts when their knowledge and expertise are needed to apply labels.

Abstract The use of intelligent machines—digital technologies that feature data‐driven forms of customization, learning, and autonomous action—is rapidly growing and will continue to impact many industries and domains. This is consequential for communities of researchers, educators, and practitioners concerned with studying, supporting, and educating information professionals. In the face of new developments in artificial intelligence (AI), the research community faces 3 questions: (a) How is AI becoming part of the world of work? (b) How is the world of work becoming part of AI? and (c) How can the information community help address this topic of Work in the Age of Intelligent Machines (WAIM)? This opinion piece considers these 3 questions by drawing on discussion from an engaging 2019 iConference workshop organized by the NSF supported WAIM research coordination network (note: https://waim.network ).

Human labeling of training data is often a time-consuming, expensive part of machine learning. In this paper, we study "batch labeling", an AI-assisted UX paradigm, that aids data labelers by allowing a single labeling action to apply to multiple records. We ran a large scale study on Mechanical Turk with 156 participants to investigate labeler-AI-batching system interaction. We investigate the efficacy of the system when compared to a single-item labeling interface (i.e., labeling one record at-a-time), and evaluate the impact of batch labeling on accuracy and time. We further investigate the impact of AI algorithm quality and its effects on the labelers' overreliance, as well as potential mechanisms for mitigating it. Our work offers implications for the design of batch labeling systems and for work practices focusing on labeler-AI-batching system interaction.

In the face of emergent external factors (e.g., supply chain disruptions or public health crises like COVID-19), businesses must adapt their business model quickly in order to ensure service continuity. However, providing recommendations regarding changes should be made to the business model is a challenging problem. First, it requires details of interactions between different components of the business (e.g., service offerings, inventory, staffing, demand) to understand what possible courses of action will have the most business impact. Second, automated models may provide recommendations on changes required in the business operations. However, with lack of human insight, it will be hard to verify the feasibility of these recommendations. Third, a generic model may not be able to provide good recommendations for diverse set of business models. Fourth, the model may not have enough features or training data to provide good recommendations.In this paper, we propose a novel approach to provide actionable items that can be recommended to business users given their business features and recommendations given to businesses in similar domain. Here we first use clustering to find the business domain and similar feature set of the domain. Then, we build a machine-learning model with explainable insights to provide recommendations on different business actions that can be taken to ensure business operations in the face of emergent external factors. Next we augment our approach with human-in-the-loop to improve its performance. Finally, we federate the machine-learning model in a similar domain to add more explainable and trusted insights and recommendations by other businesses. We describe our method, illustrate its utility with results from our implementation, and discuss areas for future work.

This note considers the topic of AI ethics as it relates to applied, industrial AI projects. In particular, it examines “intelligent sales,” a strategic management concept that envisions the enhancement of workflows within a sales organization with the use of Big Data and artificial intelligence (AI). This note examines common depictions of intelligent sales campaigns in management literature, identifying key topics in these discourses: data fusion, responsive customer care, automation, and streamlined delivery. We focus on ways in which intelligent sales is envisioned to enrich the customer-sales relationship through the use of Big Data and AI, surfacing ethical considerations around: training data and the use of AI outputs in everyday work practices. This paper contributes to discourses on the fairness, accountability, and transparency (FAccT*) of algorithmic systems by raising a number of emergent concerns in enterprise AI applications and in particular some considerations from the emergent management concept intelligent sales.

Children with autism often have sensory processing differences that can lead to a myriad of challenges, including difficulty with awareness of how their bodies occupy physical space. Natural User Interfaces (NUI) can help augment therapies to support these children. We developed DanceCraft, a whole-body interface to augment dance therapy for children with autism. In a pilot study, we deployed DanceCraft to nine homes for one week for children with autism to use. We received feedback on both the system's central activity (dance), as well as elements of its design and set up, offering lessons to guide future field deployments of such systems.

Integral to the adoption and uptake of AI systems in real-world settings is the ability for people to make sense of and evaluate such systems, a growing area of development and design efforts known as XAI (Explainable AI). Recent work has advanced the state of the art, yet a key challenge remains in understanding unique requirements that might arise when XAI systems are deployed into complex settings of use. In helping envision such requirements, this paper turns to scenario-based design, a method that anticipates and leverages scenarios of possible use early on in system development. To demonstrate the value of the scenario-based design method to XAI design, this paper presents a case study of aging-in-place monitoring. Introducing the concept of "explainability scenarios" as resources in XAI design, this paper sets out a forward-facing agenda for further attention to the emergent requirements of explainability-in-use.

Human-algorithm interaction is a growing phenomenon of interest as the use of machine learning (ML) capabilities in everyday technologies becomes more commonplace. In the workplace, such developments raise questions about how people not only make sense of algorithmic actions, but also figure out ways to collaborate with tools and systems that integrate algorithmic outputs. We draw on a field study of IT infrastructure design and report on the experiences of highly-skilled IT architects with the natural language processing (NLP) capabilities in an intelligent system under development to support their solution design work. While architects were supportive of the potential of NLP to enhance their solutioning work, they faced challenges in integrating such capabilities into their existing collaborative work practices. We discuss how these findings add nuance and complexity to discourse around the future of work.

We draw on a qualitative study of homeowners who engage in do-it-yourself (DIY) home repair and introduce "home worlds" as a conceptual lens to understand how interactions with the home as a built environment (local and mechanical engagements with its material components) are situated in broad, diffuse, and constraint-rich social worlds. A "home worlds" perspective challenges domestic computing inquiries to adopt a broader orientation when considering a home's users and requirements; our case of home repair, in particular, provokes us to consider the home as embedded in communities and framed in relation to the precarious constraints of everyday life. We discuss how these insights open up new avenues for future research and design of collaborative domestic computing technologies. ?

Engagements with social media today frequently involve the use of multiple platforms, which often exchange personal data about users. We explore individuals' experiences with the flow of personal data across platforms through interviews about dating apps, often requiring integration with a Facebook account. Inductive analysis revealed complex and, at times, conflicted ideas over what appropriate behaviors around personal data should look like in multiplatform engagements-central to these tensions were the multiple relationships that must be managed through participation in these ecosystems. For example, individuals talked of resignation in relation to specific platforms and the 'cost" of using free services (people:platform), a lack of clarity and agency in defining what personal data might be shared between platforms (people:platform:platform), and, of course, the role of personal data in their efforts to establish and maintain relationships with other users in these 'social discovery" ecosystems (people:people). We explore how these attitudes can be understood as morally charged, drawing attention to the ways in which personal data flows can enact expectations and obligations between various sociotechnical actors in personal data ecosystems and how these compound relationships reveal the complexity and texture of multiplatform participation.

Those with disabilities have long adopted, adapted, and appropriated collaborative systems to serve as assistive devices. In this paper, we present the results of a digital ethnography in a Minecraft virtual world for children with autism, specifically examining how this community has used do-it-yourself (DIY) making activities to transform the game into a variety of assistive technologies. Our results demonstrate how players and administrators "mod" the Minecraft system to support self-regulation and community engagement. This work highlights the ways in which we, as researchers concerned with accessible and equitable computing spaces, might reevaluate the scope of our inquiry, and how designers might encourage and support appropriation, enhancing users' experience and long-term adoption.

Traditional face-to-face social interactions can be challenging for individuals with autism, leading some to perceive and categorize them as less social than their typically-developing peers. Individuals with autism may even see themselves as less social relative to their peers. Online communities can provide an alternative venue for social expression, enabling different types of communication beyond face-to-face, oral interaction. Using ethnographic methods, we studied the communication ecology that has emerged around a Minecraft server for children with autism and their allies. Our analysis shows how members of this community search for, practice, and define sociality through a variety of communication channels. These findings suggest an expansion in how sociality has traditionally been conceptualized for individuals with autism.

My dissertation research explores the role technologies play in shaping how work practices are seen, imagined, and valued. I focus on how data remnants and traces, the technological residue left in the wake of human-computer interactions, become anchors that orient the construction of seeing work within an organization. To examine this, I draw on ethnographic fieldwork at a high tech firm and focus on efforts to reinvent an email client. I explore how seeing work in and through trace data paints increasingly narrow and modular portraits of work, reframing the contours and potential of vision and visibility in the workplace.

The use of online games and virtual worlds is becoming increasingly prominent, particularly in children and young adults. Parents have concerns about risks their children might encounter in these online spaces. Parents dynamically manage the boundaries between safe and unsafe spaces online through both explicit and implicit means. In this work, we use empirical data gathered from a digital ethnog-raphy of a Minecraft server, Autcraft, to explore how par-ents of children with autism continually create a "safe" virtual world through both implicit and explicit means. In par-ticular, we demonstrate how their actions in these spaces define and produce "safety," shedding light on our theoreti-cal understanding of child safety in online spaces.

Information use intrigues information behavior researchers, though many have struggled with how to conceptualize and study this phenomenon. Some work suggests that information may have social uses, hinting that information use is more complicated than previous frameworks suggest. Therefore, we use a micro‐sociological, symbolic interactionist approach to examine the use of one type of information—biomedical information—in the everyday life interactions of chronic illness patients and their families. Based on a grounded theory analysis of 60 semi‐structured interviews (30 individual patient interviews and 30 family group interviews) and observations within the family group interviews, we identify 4 categories of information use: (a) knowing my body; (b) mapping the social terrain; (c) asserting autonomy; and (d) puffing myself up. Extending previous research, the findings demonstrate use of biomedical information in interactions that construct a valued self for the patient: a person who holds authority, and who is unique and cared for. In so doing, we contribute novel insights regarding the use of information to manage social emotions such as shame, and to construct embodied knowledge that is mobilized in action to address disease‐related challenges. We thus offer an expanded conceptualization of information use that provides new directions for research and practice.