Understanding the Longitudinal Associations Between Attachment Style and AI Companion Use in Romantic Human-AI Relationships: A Three-Wave Panel Study

View at publisher
Authors: Xiaokun Yang
Cited by: 2
References: 86

Abstract

Despite research increasingly focusing on romantic relationships with AI companions, little is known about the psychological mechanisms that drive people to engage in such synthetic bonds. Based on attachment theory, this three-wave panel study first examines the reciprocal associations between attachment style and AI companion use. Random intercept cross-lagged panel models demonstrated that, at the within-person level, changes in people’s attachment anxiety over time were positively related to their AI companion use, and that, at the between-person level, people higher in attachment anxiety used AI companions more, whereas those higher in attachment avoidance used less. These findings highlight the need to disentangle within- and between-person dynamics to derive more accurate conclusions when examining underlying drivers in human-AI interactions, as well as offer new insights into the emerging human-AI relationship field by serving as a springboard for understanding how engagement with AI companions may be related to attachment styles longitudinally.

Fields of Study & Subjects

AI in Service Interactions Digital Mental Health Interventions Social Robot Interaction and HRI

Publication details

Journal / source
International Journal of Human-Computer Interaction
Publisher
Informa UK Limited
Volume
42
Issue
18
Pages
15108–15128
Publication date
28 January 2026
Language
EN