Human-AI collaboration in knowledgeintensive tasks such as fact-checking requires understanding model uncertainty in multi-document reasoning amid conflicting/agreeing evidence.Yet, existing methods only express uncertainty as numbers or hedges without revealing which evidence conflicts cause the uncertainty, leaving users unable to resolve disagreements.We present CLUE (Conflict-&Agreement-aware Language-model Uncertainty Explanations), a plug-and-play white-box framework that, to our knowledge, is the first to generate natural-language explanations of model uncertainty grounded in conflicting/agreeing evidence.CLUE (i) identifies span-level claim-evidence and inter-evidence relations that signal conflict or agreement without supervision, and (ii) uses these relations to steer explanation generation, articulating how they drive the model's uncertainty.Across three language models and two fact-checking datasets, CLUE produces explanations that more faithfully track model uncertainty and better align with the model's fact-checking decisions than span-agnostic explanation prompting; human raters also judge them more helpful, more informative, less redundant, and more logically consistent with the input.By explicitly tying uncertainty to evidence conflicts and agreements, CLUE supports practical fact-checking and other tasks that require reasoning over complex, conflicting information. * Equal contribution.Claim: Scientific data has shown that cats can be infected with SARS-CoV-2 and can spread it to other cats. ModelOutput: Supports ✅ Model Certainty: 73% [...] there is a possibility of spreading SARS-CoV-2 through domestic pets Evidence 1 [...] no further transmission events to other animals or persons Evidence 2 Automated claim verification Span interactions for model uncertainty Claim: Scientific data has shown that cats can be infected with SARS-CoV-2 and can spread it to other cats.

Fine-tuning Large Language Models (LLMs) is one response to the critique of LLMs being biased, erasing diversity, and raising ethical concerns.The Artificial Intimacy project employs artistic methods, taking personalization of chatbots to an extreme by finetuning LLMs on individual social media data.We find that regular GPT-3 chatbots attempt to circumvent value-laden content through flagging prompts and producing generic non-answers with variable success.While the transactional nature of such output allowed participants to make sense of responses with less personification, fine-tuned models presented value-laden, normative, and familiar personalities, resulting in strong personification as a way of making sense of the interactions.This mimicry of emotional connection resulted in a sense of artificial intimacy creating expectations for reciprocity and consideration that the models cannot express by design.As the commercialization of interactions with chatbots continues, we discuss the ethics of such emotional manipulation and its implications for personalization of LLMs.

A plethora of toolkits, checklists, and workshops have been developed to bridge the well-documented gap between AI ethics principles and practice. Yet little is known about effects of such interventions on practitioners. We conducted an ethnographic investigation in a major European city organization that developed and works to integrate an ethics toolkit into city operations. We find that the integration of ethics tools by technical teams desta-bilises their boundaries, roles, and mandates around responsibilities and decisions. This lead to emotional discomfort and feelings of vulnerability, which neither toolkit designers nor the organization had accounted for.We leverage the concept of moral stress to argue that this affective experience is a core challenge to the successful integration of ethics tools in technical practice. Even in this best case scenario, organisational structures were not able to deal with moral stress that resulted from attempts to implement responsible technology development practices.

The pervasiveness of large language models and generative AI in online media has amplified the need for effective automated fact-checking to assist fact-checkers in tackling the increasing volume and sophistication of misinformation. The complex nature of fact-checking demands that automated fact-checking systems provide explanations that enable fact-checkers to scrutinise their outputs. However, it is unclear how these explanations should align with the decision-making and reasoning processes of fact-checkers to be effectively integrated into their workflows. Through semi-structured interviews with fact-checking professionals, we bridge this gap by: (i) providing an account of how fact-checkers assess evidence, make decisions, and explain their processes; (ii) examining how fact-checkers use automated tools in practice; and (iii) identifying fact-checker explanation requirements for automated fact-checking tools. The findings show unmet explanation needs and identify important criteria for replicable fact-checking explanations that trace the model's reasoning path, reference specific evidence, and highlight uncertainty and information gaps.

Designing new technologies to support synchronous interaction across distances has for many years focused on creating symmetry for participation between geographically distributed actors. Symmetry in synchronous interaction has, to some extent, been achieved technologically (while multiple social, historical, political, and hierarchical concerns continue to exist) and proven empirically in the increased use of remote-work technologies that were used during the pandemic. However, synchronous interaction in hybrid work is achieved differently, since the asymmetry produced by some participants being collocated while others geographically distributed introduces increased complexities for such interactions. Focusing on this challenge, we ask: To what extent can symmetry in cooperative work engagements be achieved in hybrid work contexts? We explore this question by interrogating multiple different empirical examples of synchronous hybrid interaction collected across different organizations, activities, and events. We found that the effort required to accomplish hybrid work includes additional articulation work necessary for bounding multiple intertwined artefacts across sites, devices, and applications. Further, the multiple artefacts setup across sites, combined with asymmetric collocation across participants, produce incongruence in technological frames of reference for each participant. All participants in hybrid work have only partial access to the hybrid setup, and no single person has access to the complete setup. The incongruence in technological frames produces insurmountable gaps in collaboration, causing all hybrid work situations to be characterized fundamentally by asymmetric relationships. We argue that symmetry in hybrid synchronous interaction is impossible to attain in attempts to solve this problem through design. Instead, we propose that designers of cooperative technologies for hybrid work shift towards developing artefact-ecologies supporting hybrid work, focusing on asymmetry as a necessary feature. Fundamentally, the design strategy should explore novel ways of taking advantage of the multiple different artefact-ecologies which serve as the foundation for the hybrid collaboration. Instead of striving for symmetry, we propose to feature asymmetric conditions in future technology designs for hybrid interaction.

Sensor networks are increasingly commonplace in visions of smart cities and future healthcare systems, promising greater efficiency and increased wellbeing. However, the design of these technologies remains focused on specific users and fragmented by context, overlooking the diversity of needs, wants and values present when technologies, people, and lived realities interact within instrumented spaces. In this paper we present a workshop method – Sensing Care – that can help researchers, interdisciplinary design and development teams, and potentially affected users, to explore what it takes to design for living with sensor technologies that intersect and interact across private and public spaces, through speculative scenarios and role play. Drawing from three deployments of the workshop, we discuss how this approach supports the design of future care-oriented sensor networks, and helps designers understand what it means to live with complex technologies as people traverse diverse contexts.

One of the core goals of responsible AI development is ensuring high-quality training datasets. Many researchers have pointed to the importance of the annotation step in the creation of high-quality data, but less attention has been paid to the work that enables data annotation. We define this work as the design of ground truth schema and explore the challenges involved in the creation of datasets in the medical domain even before any annotations are made. Based on extensive work in three health-tech organisations, we describe five external and internal factors that condition medical dataset creation processes. Three external factors include regulatory constraints, the context of creation and use, and commercial and operational pressures. These factors condition medical data collection and shape the ground truth schema design. Two internal factors include epistemic differences and limits of labelling. These directly shape the design of the ground truth schema. Discussions of what constitutes high-quality data need to pay attention to the factors that shape and constrain what is possible to be created, to ensure responsible AI design.

Despite the proliferation of research on how people engage with and experience algorithmic systems, the materiality and physicality of these experiences is often overlooked. We tend to forget about bodies. The Embodying the Algorithm1 project worked with artists to explore the experience of translating algorithmically produced performance instructions through human bodies. As performers interpreted the rules of engagement produced by GPT-3, they struggled with the lack of consideration the rules showed for the limits of the human body. Performers made sense of their experience through personification, reflexivity, and interpretation, which gave rise to three modes of relating with the algorithm – agonistic, perfunctory, and agreeable. We demonstrate that collaboration with algorithmic systems is ultimately impossible as people can only relate to algorithmic systems (a one-way relation) due to the material limitations of algorithmic systems for reciprocity, understanding, and consideration for the human body.

It is not well understood why people continue to use privacy-invasive apps they consider creepy. We conducted a scenario-based study (n = 751) to investigate how the intention to use an app is influenced by affective perceptions and privacy concerns. We show that creepiness is one facet of affective discomfort, which is becoming normalized in app use. We found that affective discomfort can be negatively associated with the intention to use a privacy-invasive app. However, the influence is mitigated by other factors, including data literacy, views regarding app data practices, and ambiguity of the privacy threat. Our findings motivate a focus on affective discomfort when designing user experiences related to privacy-invasive data practices. Treating affective discomfort as a fundamental aspect of user experience requires scaling beyond the point where the thumb meets the screen and accounting for entrenched data practices and the sociotechnical landscape within which the practices are embedded.

As more and more governments adopt algorithms to support bureaucratic decision-making processes, it becomes urgent to address issues of responsible use and accountability. We examine a contested public service algorithm used in Danish job placement for assessing an individual's risk of long-term unemployment. The study takes inspiration from cooperative audits and was carried out in dialogue with the Danish unemployment services agency. Our audit investigated the practical implementation of algorithms. We find (1) a divergence between the formal documentation and the model tuning code, (2) that the algorithmic model relies on subjectivity, namely the variable which focus on the individual's self-assessment of how long it will take before they get a job, (3) that the algorithm uses the variable "origin" to determine its predictions, and (4) that the documentation neglects to consider the implications of using variables indicating personal characteristics when predicting employment outcomes. We discuss the benefits and limitations of cooperative audits in a public sector context. We specifically focus on the importance of collaboration across different public actors when investigating the use of algorithms in the algorithmic society.

“There’s an app for that” is perhaps the definitive rhetoric of our times. To understand how users navigate the trade-offs involved in using apps that support a variety of everyday activities, we conducted scenario-based semi-structured interviews (n = 25). Despite the technical and regulatory mechanisms that are supposedly meant to empower users to manage their privacy, we found that users express an overarching feeling of resignation regarding privacy matters. Because these apps provide convenience and other benefits, as one participant put it, “there is a very fine line” that marks the divide between feeling empowered in the use of technology and coping with the discomfort and creepiness arising from invasive app behavior. Participants consistently expressed being resigned to disclose data even as they accepted personal responsibility for their own privacy. We apply the findings to discuss the limits of empowerment as a design logic for privacy-oriented solutions.

These days, whether the problem is climate change or boredom, there is an app for that. The rhetoric of problem and solution, accelerated by commercial needs and salvific tech gurus, implies that software can save the world. This paper wants to start a movement/rebellion against the ubiquitous equation of P(problem) + S(software) = S(solution) as a rational approach to the ailments of this world. We question the technological effort to "playfully" afford order and control to humans through the provision of computational rules. Instead, we propose an alternative approach: designing 'pataphysical software to address familiar but ultimately imaginary problems. Defined by poet Alfred Jarry, 'pataphysics is the science of imaginary problems. Adopting the methods of 'pataphysics, we have developed mobile applications that explore invented problems and provide no solutions for them. We demonstrate how such an approach allows us to ask design questions through an aesthetic 'pataphysical practice of software development.

Designers and developers are increasingly writing manifestos to express frustration and uncertainty as they struggle to negotiate between the possibilities that IoT technologies offer, and the ethical concerns they engender. Manifestos are defining of a 'moment of crisis' and their recent proliferation indicates a desire for change. We analyze the messages manifesto authors have for their readers. Emerging from a sense of uncertainty, these manifestos create publics for debate, demand attention and call for change. While manifestos provide potential roadmaps for a better future, they also express a deep concern and even fear of the state of the world and the role of technology in it. We highlight how practitioners are responding to unstable and rapidly changing times and detail what solutions they envision, and what conflicts these might bring about. Our analysis suggests new ways HCI might theorize and design for responsibility while attending to the perils of responsibilisation.

Studying technology use in unstable and life-threatening conditions can help highlight assumptions of use built into technologies and foreground contradictions in the design of devices and services. This paper provides an account of how soldiers, volunteers, and civilians use mobile technologies in wartime, reporting on fieldwork conducted in Western Russia and Eastern Ukraine with people close to or participating directly in the armed conflict in the Donbas region. We document how private mobile phones and computers became a crucial but ambiguous infrastructure despite their lack of durability in extreme conditions of a military conflict, and their government and military surveillance potential. Our participants rely on a combination of myths and significant technical knowledge to negotiate the possibilities mobile technologies offer and the life-threatening reality of enemy surveillance they engender. We consider the problems of always-on always-connected devices under conditions of war and surveillance and our responsibilities as HCI practitioners in the design of social technologies.

Anonymization is viewed as a solution to over-exposure of personal information in a data-driven society. Yet how organizations apply anonymization techniques to data for regulatory, ethical or commercial reasons remains underexplored. We investigate how such measures are applied in organizations, asking whether anonymization practices are used, what approaches are considered practical and adequate, and how decisions are made to protect the privacy of data subjects while preserving analytical value. Our findings demonstrate that anonymization is applied to data far less pervasively than expected. Organizations that do employ anonymization often view their practices as sensitive and resort to anonymity by obscurity alongside technical means. Rather than being a purely technical question of applying the right algorithms, anonymization in practice is a complex socio-technical process that relies on multi-stakeholder collaborations. Organizational decision-making about appropriate approaches and the management of responsibility can result in workarounds necessary to negotiate the technical complexity.

Workplace health and wellness programs are increasingly integrating personal health tracking technologies, such as Fitbit and Apple Watch. Many question whether these technologies truly support employees in their pursuit of better wellness levels, raising objections about workplace surveillance and further blurring of boundaries between work and personal life. We conducted a study to understand how tracking tools are adopted in wellness programs and employees' opinions about these programs. We find that employees are generally positive about incentivized health tracking in the workplace, as it helps raise awareness of activity levels. However, there is a gap between the intentions of the programs and individual experiences and health goals. This sometimes results in confusion and creates barriers to participation. Even if this gap can be addressed, health tracking in the workplace will not be for everyone; this has implications for the design of both workplace wellness programs and tracking technologies.

Despite a proliferation of research in the use of ICTs to support active and healthy ageing, few have considered the privacy and security concerns particular to the elderly. We investigated the appropriation of tablet devices and a neighborhood portal as well as emerging privacy and security issues through ethnographic and action research in a long-term participatory design (PD) project with elderly participants. We discuss two major themes: a) the tensions related to perceived digital threats and the social pressures of online disclosure to the social environment; and b) the relation of these issues to the ICT appropriation process and the referring challenges we encountered. We argue that there is a need to understand the interleaving of physical and virtual habitats, the various ways resulting in discomfort and the senior citizens' actions -- which at first glance appear contradictory. We consider the implications of the issues observed for examining privacy and security concerns more broadly as well as discussing implications for the design of the portal and the shaping of social measures for appropriation support.

Studying in-situ technology use over time can be difficult and this is especially so when considering technologies such as activity tracking devices explicitly designed to be unobtrusive. Yet understanding activity tracking in practice is crucial, as tracking technologies become important tools for health promotion and health insurance programs. In this paper, we describe a method for a longitudinal participant-driven photo elicitation study of activity tracking. During the five-month long study, our drop-out rates were low and we observed idiosyncratic practices with lapses and particular use patterns among participants along with significant self-reflection on activity tracking as a practice. We describe our method in detail, discussing the necessary adaptations for the study of activity tracking practices. We offer our experiences of benefits and challenges of this process, and suggest points for consideration for future studies in the area.

Personal health technologies are increasingly introduced in workplace settings. Yet little is known about workplace implementations of activity tracker use and the kind of experiences and concerns employees might have when engaging with these technologies in practice. We report on an observational study of a Danish workplace participating in a step counting campaign. We find that concerns of employees who choose to participate and those who choose not to differ. Moreover, privacy concerns of participants develop and change over time. Our findings challenge the assumption that consumers are becoming more comfortable with perceived risks associated with wearable technologies, instead showing how users can be initially influenced by the strong positive rhetoric surrounding these devices, only to be surprised by the necessity to renegotiate boundaries of disclosure in practice.

Sedentary work is a contributing factor to growing obesity levels worldwide. Research shows that step-counters can offer a way to motivate greater physical mobility. We present an in-situ study of a nation-wide workplace step-counting campaign. Our findings show that in the context of the workplace steps are a socially negotiated quantity and that participation in the campaign has an impact on those who volunteer to participate and those who opt-out. We highlight that specific health promotion initiatives do not operate in a vacuum, but are experienced as one out of many efforts offered to the employees. Using a social ecology lens we illustrate how conceptualizing a step-counting campaign as a health promotion rather than a behavior change effort can have implications for what is construed as success.

HCI research has explored mobile technologies to support social activity and to support greater feelings of connectedness. Much of this has focused on different mobile devices, individual preferences and modes of use. Yet social activity and connectedness are about ongoing enactments of relationships across technologies. We propose the relational lens as a way to include a notion of relational tension in addition to individual preferences in the design and analysis of mobile communication technologies. We discuss three strategies people use to manage tensions in their relationships: selection, segmentation and integration. Our data show that use of social technologies can at times destabilize social relations and occasion relational tensions, forcing users to renegotiate how they enact these relationships.

One of the most significant contemporary technological trends is institutional adoption and use of mobile and location‐based systems and services. We argue that the notion of “location” as it manifests itself in location‐based systems is being produced as an object of exchange. Here we are specifically concerned with what happens to institutional roles, power relationships, and decision‐making processes when a particular type of information—that of spatiotemporal location of people—is made into a technologically tradable object through the use of location‐based systems. We examine the introduction of GPS (Global Positioning Systems) technologies by the California criminal justice system and the institution of parole for monitoring the movements of parolees, with consequences both for the everyday lives of these parolees and the work practices of their parole officers. We document the ways in which broad adoption of location‐based and mobile technologies has the capacity to radically reconfigure the spatiotemporal arrangement of institutional processes. The presence of digital location traces creates new forms of institutional accountability, facilitates a shift in the understood relation between location and action, and necessitates new models of interpretation and sense making in practice.

Mobile devices are playing an increasingly intimate role in everyday life. However, users can be surprised when informed of the data collection and distribution activities of apps they install. We report on two studies of smartphone users in western European countries, in which users were confronted with app behaviors and their reactions assessed. Users felt their personal space had been violated in "creepy" ways. Using Altman's notions of personal space and territoriality, and Nissenbaum's theory of contextual integrity, we account for these emotional reactions and suggest that they point to important underlying issues, even when users continue using apps they find creepy.

It is not surprising that HCI researchers are attracted to the role of technology in global processes as many of us already live inherently transnational lives. While the notion of global connectedness is hardly new, the issues that confront us are more than specific concerns for remote migration, distributed work, or developing nations. Rather, we argue that transnational HCI is a contemporary condition of the design and use of technological systems, both at home and abroad. This special issue of Human-Computer Interaction is dedicated to exploring how and why a transnational lens matters to the study, design, and development of computational systems. We consider this theoretical perspective in terms of both present technology use to construct and manage transnational relations and processes, and the possibilities such a lens opens for future research and design. The papers in this issue contribute to the field of HCI by bringing the principles developed in anthropology, sociology, and elsewhere to bear on the conversation in HCI, retooling them for our present context, while preserving the richness of their methodological orientation.

In this paper we report on the design and implementation of an initial prototype to explore how to better situate in the home social media content individually generated by family members. We considered whether existing infrastructure and practices of social media might be leveraged to offer new kinds of shared family experiences. We found that families perceived the system to be "cosy" and intimate, especially in contrast to Facebook, and as a result 'shared to care'. While aspects of the design had a strong role to play in faciliating this perception, participants enacted their own boundaries of sharing and disclosure based on pre-existing practices and attitudes toward social technologies. The study demonstrated that there are productive design opportunities in home systems that can leverage content via a broad range of social media applications.

While social network capabilities are proliferating on many online services, research has focused on just a few popular social network sites. In this note, we consider a different kind of social network site, explicitly designed to support particular types of risky sexual activity among men who have sex with men (MSM). We consider the role of ambiguity built into the interface in how users manage self-disclosure and its association with articulating more friends-only or sexual connections on the site. Despite the site's explicit orientation toward risky sexual practices, we find indications that users mitigate potential public health issues through the practice of sero-sorting. We discuss how design considerations that may allow for easier entrance into a community can cause problems for long-term users, or generate potential public health issues.

In this article we describe people's online contribution practices in contexts in which the government actively blocks access to or censors the Internet. We argue that people experience blocking as confusing, as a motivation for self-censorship online, as a cause of impoverishment of available content and as a real threat of personal persecution. Challenging ideas of blocking as a monolithic, abstract policy, we discuss five strategies with which Internet users navigate blocking: self-censorship, cultivating technical savvy, reliance on social ties to relay blocked content, use of already blocked sites for content production as a form of protection and practiced transparency. We also discuss strategies that forum owners and blogging platform providers employ to deal with and to avoid blocking. We conclude by advocating for more research that acknowledges the complexity of the contexts in which all Internet users contribute to the Internet and social media.

Location-based ubiquitous computing systems are entering mainstream society and becoming familiar parts of everyday life. However, the settings in which they are deployed are already suffused with complex social dynamics. We report on a study of parole officers and parolees whose relationships are being transformed by location-based technologies. While parolees are clearly subjects of state discipline, the parole officers also find themselves subject to new responsibilities. This study highlights the complexities of power in sociotechnical systems and what happens when location becomes a tradable, technological object.

Many observers have praised new communication technologies for providing convenient and affordable tools for maintaining relationships at a distance. Yet the precise role of mediated communication in relationship maintenance has been difficult to isolate. In this paper, we treat residential moves as natural experiments that threaten existing social relationships and often force people to rely on mediated communication to maintain their old relationships. Results from a 3-wave survey of 900 residential movers describing 1892 relationships shows that email and the telephone play different roles in social relationships. Email helps maintain social relationships, in the sense that relationships decline when email drops after the move. However increases in email are not associated with increases in the depth of the relationship or exchanges of support. In contrast, phone calls help movers grow relationships and exchange social support.

Disasters affect not only the welfare of individuals and family groups, but also the well-being of communities, and can serve as a catalyst for innovative uses of information and communication technology (ICT). In this paper, we present evidence of ICT use for re-orientation toward the community and for the production of public goods in the form of information dissemination during disasters. Results from this study of information seeking practices by members of the public during the October 2007 Southern California wildfires suggest that ICT use provides a means for communicating community-relevant information especially when members become geographically dispersed, leveraging and even building community resources in the process. In the presence of pervasive ICT, people are developing new practices for emergency response by using ICT to address problems that arise from information dearth and geographical dispersion. In doing so, they find community by reconnecting with others who share their concern for the locale threatened by the hazard.

In this paper we examine how routine uses of the Internet for communication with family and friends and for entertainment may serve as indicators of overall levels of psychological well-being. Changes in psychological well-being in response to a major life event, such as a residential move, can drive changes in routine uses of the Internet, suggesting Internet-based coping strategies. Specifically, women who report high levels of depressive affect, decrease internet use for communication. Men with similar levels of depressive affect increase internet use for entertainment. We discuss implications of these findings for our understanding of the role of the Internet in everyday behavior and instances of coping with stressful situations.

One of the outcomes of massive adoption of technology is that much of daily technology use and consumption is embedded into "unremarkable" daily life routines. Occasionally, these routines undergo major shifts, often in conjunction with major life events such as marriage, birth of a child, or a residential move. We propose a model of settling into a new location as a function of balance between the pull of the things left behind and the demands of the new and unknown. It is through this experience of being unsettled that we explore the processes of behavior adjustment and re-evaluation of old patterns of technology use as it relates to the old location and the demands of the new location.