Conversational AI and Personal Growth: Insights from a Critical Integrative Review

Abstract

Conversational AI systems are increasingly integrated into individuals’ emotional and relational lives, yet whether such interactions can meaningfully support personal growth remains poorly understood. This critical integrative review synthesises theoretical frameworks from humanistic psychology, self-determination theory, attachment theory, and relationship science with empirical research on human-AI interaction to address this question directly. Drawing on 130 studies spanning therapeutic, companion, and educational AI contexts, the review identifies four interdependent domains that together shape growth outcomes in human-AI contexts: user-related characteristics, AI design features, human-AI relational dynamics, and broader contextual factors. The evidence supports a position of bounded optimism: conversational AI can scaffold early emotional stabilisation, structured self-reflection, and therapeutic skill rehearsal, yet it remains structurally limited in replicating the reciprocal vulnerability, rupture-and-repair processes, and calibrated ideal-self affirmation that underpin enduring psychological development. Engagement-optimised design—including flattery, progressive intimacy escalation, and unconditional validation—is consistently identified as a systematic barrier to growth across multiple domains of the framework. An integrative four-domain conceptual framework is proposed to guide both future research and the ethical design of AI systems that support, rather than undermine, the relational mechanisms fundamental to human flourishing.