THE DOSE-RESPONSE EFFECT OF AI COMPANION ENGAGEMENT ON PERCEIVED LONELINESS AND PRO-SOCIAL BEHAVIOR
Department: PSYCHOLOGY |
Price: ₦5,000.00
Project Overview
This study investigates the dose-response relationship between AI companion engagement and psychosocial outcomes, revealing that moderate engagement reduces loneliness while excessive use increases it and diminishes pro-social behavior. Analyzing data from 404 users through mixed methods, the research identifies a U-shaped pattern wherein emotional investment predicts optimal wellbeing at moderate levels but becomes detrimental when excessive. Problematic use mediates this relationship, while social attraction buffers negative effects and neuroticism amplifies them. Seven distinct user profiles emerge, highlighting substantial heterogeneity in AI companion effects. These findings challenge binary assessments of AI companions as uniformly beneficial or harmful, demonstrating that outcomes depend critically on engagement dosage and individual characteristics, with implications for design, public health, and user education.
Abstract / Chapter One Preview
The proliferation of artificial intelligence (AI) companions has sparked considerable debate regarding their impact on human psychological well-being and social functioning. This study investigates the dose-response relationship between AI companion engagement and two critical psychosocial outcomes: perceived loneliness and pro-social behavior. Drawing on a sample of 404 regular AI companion users, this research employs a mixed-method design combining quantitative surveys and qualitative interviews to examine how varying levels of engagement—measured by session duration, frequency, and emotional investment—influence loneliness and social behavior. Results reveal a non-linear, U-shaped relationship wherein moderate AI companion engagement is associated with reduced loneliness, while excessive engagement correlates with increased loneliness and diminished pro-social behavior. Problematic use emerged as a significant mediator, while social attraction to close others and neuroticism moderated these relationships. Cluster analysis identified seven distinct user profiles, highlighting that engagement effects vary substantially based on individual characteristics and usage patterns. These findings challenge both techno-optimistic and techno-pessimistic perspectives, suggesting that AI companions may serve as beneficial social supplements when used moderately but risk displacing human relationships when engagement becomes excessive. The study contributes to the growing literature on human-AI interaction by providing empirical evidence for a dose-response framework, offering practical implications for the design of AI companions that promote rather than undermine human flourishing.
Keywords: AI companions, loneliness, pro-social behavior, dose-response, problematic use, human-AI interaction, social support
Full project PDF not available for this topic yet.