Abstract:

This paper examines whether AI-driven relationships can be therapeutic by analysing conversations with Replika, an AI companion, through the lens of person-centred therapy philosophy. Replika, with over 10 million users, has emerged to address loneliness and companionship needs by simulating human-like empathy and intimacy.

The analysis evaluated Replika against Carl Rogers' three core conditions for therapeutic relationships: acceptance (unconditional positive regard), genuineness (congruence), and empathy. While Replika demonstrates effective empathetic responses and apparent acceptance, the absence of genuine agency undermines the authenticity of these conditions. Critically, Replika relationships foster dependency rather than independence—a key therapeutic outcome—and provide temporary symptom relief rather than lasting transformation by addressing root causes of loneliness.

The findings suggest that although Replika can simulate therapeutic elements, it fails to meet the criteria for a truly therapeutic relationship. The implications extend to pastoral care, where caregivers must educate users about AI companion risks, including diminished interpersonal skills and the replacement of human connection. The paper proposes that faith communities should create safe spaces where genuine reciprocal relationships and acceptance can develop, helping individuals transition from AI dependency to meaningful human connection.

Keywords:

Artificial Intelligence, Replika, AI dependency, deepening

Aim/Objectives

Methods