Answers are grounded in this paper and cite their sources.
This systematic review analyzes 47 studies on autonomous socially interactive agents (SIAs) designed for long-term use in well-being contexts. Drawing on literature from the past 7 years, we examine how these systems are built to establish and sustain meaningful relationships with users over time. The analysis reveals three central trends: the increasing use of personalization strategies, the adoption of hybrid or data-driven architectures to support flexible interaction, and a growing emphasis on ethical design considerations such as transparency and user control. We highlight how these design choices contribute to the relational quality of human–agent interactions and identify gaps in current approaches to memory, adaptivity, and relational flexibility. The findings suggest that successful long-term SIAs must move beyond task completion to support evolving, trust-based relationships that are responsive to users’ needs, emotions, and preferences. We conclude by outlining key research opportunities for developing agents that are not only effective but also relationally and ethically attuned.
No takes yet. Share an insight, caveat, or question.
Antunes et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: