From Adolescence to Older Adulthood: Lifespan Pathways Linking AI Companion Chatbots to Mental Health

ABSTRACT: AI-based conversational agents are increasingly used for emotional support, companionship, and day-to-day coping. These systems can provide immediate reassurance, reduce distress in the moment, and offer a low-barrier channel for reflection. At the same time, concerns are growing that frequent reliance on AI companions may displace human relationships and narrow users’ exposure to the interpersonal friction that supports psychological growth. This narrative review synthesizes conceptual and empirical themes to explain how AI companion chatbot use may relate to loneliness and depressive symptoms across the lifespan. We propose a developmental framework distinguishing supportive pathways (e.g., perceived availability, emotion regulation scaffolding, and social activation) from risk pathways (e.g., social displacement, dependency, avoidance coping, and affirmation-biased feedback loops). A central contribution is a lifespan account of how positive-only or preference-aligned feedback may undermine constructive stress appraisal, frustration tolerance, resilience, and grit—capacities that are built through repeated experiences of manageable challenge, honest feedback, and relationship repair. We conclude with implications for practice, education, and design, emphasizing developmental tailoring, safeguards against over-reliance, and research priorities needed to clarify causal mechanisms and long-term outcomes.