Artificial companionship has emerged as a prominent and rapidly evolving application domain of artificial intelligence (AI). AI companions — generative conversational chatbots, avatars, and related systems designed or used to simulate human-like relationships — are increasingly adopted as emotional support tools, wellness coaches, study partners, language tutors, customer engagement agents, therapeutic assistants, and digitally mediated confidants (Chen et al., 2024; Ciriello et al., 2025a; Peter et al., 2025). Across both societal and business contexts, these systems are designed to engage users through conversation, memory, personalisation, and simulated responsiveness, thereby entering domains historically shaped by human empathy, trust, and care (Montemayor et al., 2022).
Unlike conventional productivity tools, AI companions derive much of their potential value from sustained interaction, perceived attentiveness, inferred preferences, and personalised continuity over time. Their effects therefore unfold relationally rather than through discrete task outputs alone, rendering constructs such as trust, dependency, wellbeing, dignity, identity, and power analytically central to understanding their design, use, and governance. Understanding such systems requires moving beyond purely information-processing views of intelligence toward richer accounts that incorporate emotion, embodiment, and situated human experience (Stein & Shollo, 2025). At the same time, although debates surrounding machine consciousness remain ongoing (Breidbach et al., 2025), these systems do not presently possess human-like agency or understanding. Rather, they simulate relational behaviour through statistical pattern generation. Consequently, the relational effects they generate should be understood as sociotechnical phenomena emerging from interactions among users, systems, and contexts, rather than as properties of the technology alone (Riemer et al., 2026).
For Information Systems (IS) scholarship, this development invites a more explicit engagement with relational aspects of emerging digital technologies. Much prior IS research has understandably emphasised efficiency, adoption, performance, and managerial outcomes (Clarke & Davison, 2020). While these remain important, AI companions call for complementary perspectives that examine how technologies shape social relations, emotional experiences, and quality of life. A relational perspective does not replace instrumental concerns, but extends them by recognising that systems designed for sustained interaction may reconfigure relationships among individuals, organisations, markets, and institutions. These questions are especially salient for IS scholars, given the field’s long-standing concern with the design, governance, and use of sociotechnical systems (Sarker et al., 2019). This builds on prior IS work that conceptualised human–technology interaction in relational terms, for example through constructs such as social presence and perceived social relationships with technological agents (Qiu & Benbasat, 2009).
Recent advances in generative AI reinforce the importance of this shift by transforming digital platforms through intelligent automation, democratisation, hyper-personalisation, and collaborative value co-creation, thereby reshaping relationships among platform stakeholders (Wessel et al., 2025). AI companions may extend these dynamics by shifting personalisation from content-level adaptation to interaction-level continuity, enabling systems to adapt not only what they produce, but also how they moderate users’ feelings. Generative AI is increasingly positioned not merely as a tool, but as a collaborator, coach, or companion in personal and organisational contexts, with emerging research suggesting that such interactions may reshape cognition, identity, and capabilities by influencing creativity, decision-making, emotional regulation, and self-perception (Adam et al., 2025). These developments raise fundamental questions about whether reliance on AI companions augments or erodes human relational capacities, and how individuals adapt to increasingly personalised and emotionally responsive digital environments (Chen et al., 2024). AI companions thus cannot be understood solely in terms of technical capabilities, as their outcomes are co-produced through interface design, data architectures, platform incentives, governance arrangements, user practices, and broader social contexts, while also reflecting platform-based business models that shape how relational value is created, captured, and distributed (Arsel, 2025; Peter et al., 2025; Yoo et al., 2024)
Early evidence points to sharply double-edged implications. On the one hand, AI companions may broaden access to personalised support, augment overstretched educational and care services, and improve continuity in customer and employee assistance. On the other hand, they may intensify surveillance, emotional manipulation, behavioural influence, commodified intimacy, and new forms of dependency (Chen et al., 2024; Ciriello et al., 2025a). Commercial systems may generate value precisely from moments of loneliness, uncertainty, or vulnerability, raising important ethical concerns. Organisational deployments of relational AI may similarly enhance service quality while reshaping labour processes, authority structures, accountability, and professional boundariess (Ciriello et al., 2025a; Seeber et al., 2020). More fundamentally, these dynamics may exceed simple risk–benefit trade-offs. Recent research suggests that AI companionship constitutes a sociotechnical paradox in which users form emotionally meaningful bonds while confronting AI’s lack of genuine reciprocity, producing tensions between attachment and detachment, dependency and self-estrangement (Ciriello et al., 2025b). These tensions warrant careful inquiry that moves beyond purely technocratic or narrowly managerial accounts.
This special issue seeks to advance a focused IS conversation on the evolving role of AI systems designed not only to support tasks, but also to sustain interaction, offer guidance, personalise engagement, and participate in socially embedded relationships. To fit the special issue, manuscripts should clearly align with the CfP’s definition of AI companions or closely related relational AI systems and explain how the study advances an IS conversation on the design, governance, or use of sociotechnical human–AI systems. Manuscripts on general GenAI, AI ethics, human–AI interaction, or digital platforms are in scope only where they foreground a clear relational or companionship component.
We invite theoretically ambitious and empirically grounded research on the individual, organisational, and societal implications of AI companions and related systems (e.g., AI co-pilots, coaches, tutors, and colleagues in contexts involving emotional or relational engagement). We particularly welcome interdisciplinary and critical perspectives, alongside behavioural and design-oriented work, that broaden current understandings of how relational AI systems are conceptualised, designed, governed, and used. We encourage submissions that engage deeply with the nature of generative AI technologies and avoid treating them as "just another IT artefact" or uncritically adopting anthropomorphic assumptions (Riemer et al., 2026). Instead, we seek work that carefully theorises the distinct sociotechnical characteristics of these systems and their implications for IS theory and practice.
We welcome diverse methodologies, including qualitative, quantitative, mixed-method, experimental, ethnographic, computational, historical, and design science research. Empirical submissions should be grounded in rigorous and credible data and methods appropriate to their claims. The special issue has a strong preference for empirically grounded work. Purely conceptual submissions may be considered on a case-by-case basis, but must clear an exceptionally high bar by offering a clearly novel and potentially field-shaping theoretical contribution, rather than broad commentary, literature review, or early-stage theorising. Submissions should make a clear IS contribution by advancing understanding of the design, governance, or use of these sociotechnical systems.