Exploring Attachment and Trust in AI Companion Use

This ongoing study explores how individual attachment styles influence AI companion usage and how interpersonal trust moderates this relationship. By examining these psychological factors, we aim to understand the underlying mechanisms that drive user engagement with AI companions. As these companions become increasingly human-like and integrated into users' personal lives, insights from this research are critical for informing the ethical design and governance of such technologies. Drawing on attachment theory, we hypothesize that users with different attachment styles will exhibit distinct patterns of AI companion use, and that interpersonal trust will significantly affect these patterns. A survey using standardized scales will collect quantitative data from experienced AI companion users, with structural equation modelling (SEM) employed to analyze the relationships between variables. The findings are expected to provide theoretical contributions by extending attachment theory into human-AI interactions. Practically, the study will offer actionable insights for AI designers and policymakers to develop companions that are responsive to users' psychological needs, promoting well-being and minimizing risks like over-dependency. By aligning AI companion design with users' attachment styles and trust dynamics, we aim to enhance the ethical alignment of these tools with users' needs and behaviors.