Umeå University's logo

If a technological agent treats people unequally, it may be perceived as “being unfair.” But in what sense can fairness be considered an attribute of a nonhuman entity – a thing? This paper addresses this question through an exploratory study combining an experiment and a focus group. In the experiment, implemented as a quiz game hosted by an agent, two levels of participants’ Treatment by the agent (Fair/Unfair) were combined with two levels of agents’ Anthropomorphism (High/Low). Data about participants’ perceptions of the agents were collected through Likert scales and post-session interviews. A subset of participants took part in a follow-up focus group study, in which they shared their thoughts and reflections on intelligent agents’ fairness, grounded in their prior quiz game experience. The results suggest that while perceived fairness of an agent is a key aspect of human-agent interaction, operationalizing it is complicated by its ambiguity, context dependence, and entanglement with other aspects of interaction.

This paper highlights and discusses the methodological challenge of combining control and validity in social robotics research. Proactive, future-oriented studies that explore novel technologies and interactions are usually conducted in specially constructed environments and involve completing researcher-defined tasks. While researchers’ control over environments and tasks allows for investigating human-robot interaction phenomena that do not (yet) exist in real-life settings, it may also potentially undermine the validity of the studies. The paper discusses combining control and validity in relation to existing social robotics research and identifies key topics for further methodological developments. The paper argues that detailed and explicit representations of study contexts are crucial to ensure creating consistent study environments. It is also argued that there is a need for analytical tools supporting an understanding of how study participants frame technologies and tasks within particular study contexts.

This study presents KindMind, a conversational tutoring robot designed to investigate how politeness and fairness can be effectively embedded in AI-driven educational systems. Focusing on English grammar instruction, the project explores user perceptions of respectful and transparent chatbot behavior and translates these expectations into actionable design principles. A mixed-methods approach was adopted, involving 22 survey participants and five semi-structured interviews with learners and educators aged 18 to 58. Results show that users associate politeness with patient phrasing, encouragement, and a calm tone, while fairness is linked to clarity, equal opportunity, and non-judgmental feedback. These insights informed the development of structured prompt strategies, accessible user interfaces, and emotionally intelligent response flows. KindMind was evaluated through user testing with scripted interactions and survey feedback, confirming its ability to deliver a respectful and trustworthy user experience. By grounding interaction design in ethical communication principles, the system offers a replicable model for inclusive and learner-centered conversational AI.

While global stress levels have been increasing, there has also been a notable rise in the use of stress management applications. These applications commonly monitor the user's stress levels and visualize this information or offer digital versions of evidence-based methods for stress relief. Additionally, recent advances in artificial intelligence provide significant potential to enhance stress management applications through increased personalization and the facilitation of social interaction. In many regards, these applications are "user-friendly,"but there seems to have been little effort to include users in the first stages of development. We believe that by including potential users early in the development process, promising designs can be uncovered that lead to a better appropriation of the application into users' daily lives. In this paper, we address the research gap by exploring how potential users anticipate using an intelligent stress management application. We achieve this by allowing seven participants (n = 7) to imagine having access to such an app for two weeks, and then conducting interviews with them about their experience. Our main findings are that a) the role of an intelligent stress management application is multi-faceted, b) the social interaction should be friendly and feel genuine, c) there is a general need for task management and scheduling that takes stress levels and energy into consideration and d) designing for smartphone is challenging because it is often a source of stress, but an intelligent system can decrease such stress.

Social robots and other technologies designed to engage in social interactions with their users increasingly enter our everyday environments. Dealing with conceptual and practical challenges, related to these developments, requires an understanding of these technologies as new types of social actors. However, despite the substantial research effort in HRI and related fields, many basic questions about the meaning of 'social actors' in the case of technological artifacts largely remain open. In this paper, we argue that a potential way to address the questions is avoiding essentialist assumptions about the nature of technological artifacts as social actors. Instead, we propose focusing on contextual framing of artifacts, that is, on how people perceive the artifacts as embedded in meaningful social contexts. In line with this proposition, we call for the development of conceptual and methodological tools that allow researchers to systematically address contextual framing. To illustrate potential directions for this development, we propose tentative versions of two analytical tools: (1) a conceptual model of robot-mediated collaboration, and (2) a set of analytical dimensions specifically focusing on contextual framing of technological agents.

As Artificially Intelligent systems are becoming more and more present in our surroundings, our ways of interacting with them are also changing. From commercial chatbots to home assistants and robot companions, machines are progressively taking up the role of "communicators", provided with their own agency, and able to interact with their human counterparts in new ways. This workshop aimed at gathering experts in fields relevant to the study of AI systems as communicators, including but not limited to Human-Computer Interaction, Artificial Intelligence, Human-Robot and Human-AI Interaction. It was organized in order to discuss new challenges brought by this recent shift, compare methods and perspectives between different fields, and foster long-term collaborations.

The focus of this chapter is on the perspectives on human actions that were adopted in HCI and on how these perspectives influenced HCI research. It is argued that, historically, particular views on human actions played a major role in determining the specific research agendas, methodologies, and outcomes of HCI studies. The discussion in this chapter follows a general trend in HCI research toward increasingly more advanced notions on human actions and a respective series of extensions of the scope of the field from users' tasks to meaningful actions in real-life contexts. The implications of the broadened scope of HCI for conceptual analyses, empirical studies, and design explorations are discussed. This chapter concludes with reflections on new conceptual developments, addressing transformative and emancipatory effects of technology on human, and even non-human, beings.

Technology-supported stress management is one of the most promising and practically important application areas for intelligent companions, that is, digital assistants evoking empathy and personal attachment. In this paper, we employ the method of Futuristic Autobiographies (FABs) to elicit participants' attitudes and reflections regarding an imaginary digital assistant for stress management, implemented as an intelligent companion. For the purposes of the study, we developed six FABs, highlighting a range of potential issues related to the use of intelligent companions. The participants (N=17) provided their responses to the FABs by completing a survey, and a subset of the participants (N=5) also took part in follow-up online interviews. A thematic analysis of the results revealed six main themes: objectivity of the digital assistant, human-likeness, context-specificity of assistant's behavior, user's control, stress management, and user privacy. The themes, as well as the implications of the results for the design and use of intelligent stress management companions, are discussed.