UC Riverside

This narrative review examines research on artificial intelligence (AI), including rule-based systems, natural language processing models, and large language models, in relation to well-being, social connection, and personal growth. After briefly tracing the history of AI, we review evidence from AI-facilitated well-being interventions, educational applications, interpersonal skill development, AI-mediated communication, and AI companionship. In clinical and nonclinical settings, structured AI applications show some short-term benefits for anxiety, stress, loneliness, self-esteem, learning, and social confidence, while emerging evidence suggests that AI companions may provide temporary emotional support and a sense of connection. However, findings across these domains are not consistent and appear to depend on how AI is used, the structure of the interaction, the type of feedback provided, and the broader context. Important risks include emotional dependence, overreliance, reduced human connection, weakened authenticity in communication, cognitive or socioemotional skill erosion, bias, and poor crisis response. Preliminary findings suggest that AI may be most beneficial when used to support, rather than replace, human capacities and relationships. Future research should examine long-term outcomes, individual differences, real-world use of publicly available AI systems, and the conditions under which AI strengthens or undermines well-being, relationships, and personal growth.