As artificial intelligence (AI)-based robots become increasingly embedded in daily life, their potential to support human well-being has attracted growing interest from both academia and mass media. Based on their embodiment and presence, AI-based robots can be divided into physical robots and virtual robots. Although research on AI-based robots is expanding, it remains fragmented across disciplines and lacks a cohesive understanding of how users interact with AI-based robots in real-world contexts and how such interactions influence individuals’ well-being. Previous studies have provided promising evidence for the effectiveness of AI-based physical and virtual robots in supporting personal well-being, particularly by offering emotional support, reducing loneliness, and fostering social connections. However, these studies have largely been conducted in controlled or structured settings and tend to conceptualize human–robot interaction (HRI) as a short-term intervention. Critical gaps remain in understanding the dynamic, evolving, and co-constructed nature of HRI and the resultant outcomes for multiple dimensions of individual well-being in daily life.

In order to address these gaps, this dissertation aims to investigate how human interactions with AI-based robots contribute to individual well-being in daily life. The dissertation comprises five publications: two systematic literature reviews (Publications I and II) and three empirical case studies (Publications III-V).

Publications I and II conduct two systematic literature reviews (SLRs) that examine existing research on the use of AI-based physical and virtual robots in healthcare and daily settings, respectively. The findings provide a conceptual foundation and source of questions for subsequent empirical research through synthesizing current knowledge on robot designs, application contexts, target users, and the antecedents and consequences of robot use. Moreover, a clear research agenda for future investigation in the field is proposed, including diversifying robot types, application contexts, and stakeholders; advancing research methodologies; expanding thematic dimensions; and embracing multidisciplinary perspectives.

This dissertation further conducts empirical studies (Publications III-V) that examine individual interactions with AI-based physical and virtual robots for well-being support in real-life scenarios. Adopting a multiple-case study approach, these publications explore five distinct cases involving both physical and virtual robots across different daily life settings.

Specifically, Publication III conducts a single case study on user-generated reviews of Robot A from social media platforms. Through an inductive thematic analysis approach, it explores the embodied interactions between users and an AI-based physical robot in home environments and their positive impacts on individual well-being, drawing upon the HRI studies. Likewise, Publication IV conducts a single case study on tweet data from Robot B, an AI-based virtual robot, in digital social environments. A mixed-methods approach that combines quantitative topic modeling with qualitative interpretation is employed to uncover the relationship dynamics between users and AI-based virtual robots, guided by social penetration theory. Drawing on actor-network theory, Publication V conducts a multiple case study on three AI-based virtual robots for mental healthcare (Robots C, D, and E) and explores how users and AI-based virtual robots interact as focal actors to cope with individuals’ emotional issues from a sociotechnical perspective. This study applies a mixed-methods approach, combining quantitative emotion analysis with qualitative thematic analysis.

The empirical results of this dissertation reveal that the interaction between users and AI-based robots is a highly contextual, dynamic, and co-constructed process, shaped by robot embodiment, functionality, contexts, and user needs. These interactions follow nonlinear trajectories, while users may delay or withdraw from interactions due to various negative experiences. Additionally, this dissertation provides empirical evidence of the positive well-being outcomes of HRI on individual well-being across emotional, social, cognitive, and behavioral dimensions in daily life. Notably, these well-being outcomes are not only driven by robot features or technical performance but also emerge from the sustained and stable interactions between users and AI-based robots.

This dissertation contributes to the academic fields of Information Systems and HRI. First, it systematically synthesizes and conceptualizes the current state of research on AI-based physical and virtual robots in well-being contexts, offering a structured theoretical foundation for understanding this emerging area. Second, this dissertation advances a contextualized and dynamic process-oriented understanding of HRI by demonstrating how robot representation (physical vs. virtual), functional design (companionship vs. therapy), and user needs (emotional, social, cognitive) collectively shape and develop interaction patterns. Third, this dissertation responds to growing calls in the healthcare digitalization field to move beyond functional or performance-based evaluations, offering a more holistic view of how AI-based robots become integrated into users’ emotional, social, cognitive, and behavioral well-being. From a practical perspective, this dissertation provides actionable insights for stakeholders involved in the design, deployment, and governance of AI-based robots in well-being support. The findings contribute to bridging the gap between technological innovation and social responsibility, advancing the ethical, inclusive, and meaningful integration of AI-based robots into our society.