Are Measures of Children's Parasocial Relationships Ready for Conversational AI?

Abstract

Parasocial scales assume one-sided bonds with media figures—but what happens when the media talks back? The rapid emergence of conversational AI has transformed how children interact with digital characters, yet existing parasocial interaction (PSI) and relationship (PSR) measures are still grounded in passive television viewing. Through a systematic review of 24 studies spanning 1976–2024, we reveal critical gaps in current measurement approaches for understanding child-AI interactions. Our analysis exposes methodological weaknesses, including age-indiscriminate measurement tools, inconsistent conceptualisation of character realism, and an oversimplified positive friendship framework. Existing measures predominantly treat PSRs as unidirectional, static constructs, failing to capture the long-term, reciprocal nature of AI interactions. We identify promising directions for measurement evolution, including the revival of early affective scales, refined behavioural measures, and an expanded relationship taxonomy. This analysis provides an urgent review of current frameworks to help us understand and assess children’s relationships with increasingly sophisticated AI characters.