Loneliness, Social Connectedness, and Conversational AI Use among Kagay-anon Youth: A Mixed-Methods Study

This study examined the relationships among conversational AI, loneliness, social connectedness, and its use for emotional support among Kagay-anon youth, focusing on its use as a source of emotional support. An explanatory sequential mixed-methods design was employed, combining quantitative surveys using the modified revised UCLA Loneliness Scale and Chatbot Usability Scale with qualitative semi-structured interviews. Quantitative data were analyzed using descriptive statistics and Pearson correlation, while qualitative data underwent thematic analysis. The quantitative findings indicated that loneliness was positively associated with conversational AI use for emotional support (r = .42, p < .001), whereas social connectedness was negatively associated with conversational AI use for emotional support (r = –.36, p < .001). Interviews further explained these findings by showing that participants turned to conversational AI during periods of emotional distress, valued its accessibility, nonjudgemental responses, and opportunities for perceived emotional expression and self-reflection, while also recognizing its limited emotional depth and the potential risk of overreliance. Five key themes emerged: (1) emotional struggles leading to conversational AI use, (2) barriers to seeking human support, (3) conversational AI as a source of emotional comfort and reflection, (4) accessibility and convenience driving conversational AI use, and (5) conversational AI's perceived limitations and concerns about emotional dependency. Among 437 Kagay-anon youth, loneliness was positively associated with conversational-AI use for emotional support (r = .42, p < .001), while social connectedness was inversely associated with such use (r = −.36, p < .001). Follow-up interviews with 13 informants explained these patterns through emotional struggles, barriers to human disclosure, perceived comfort and reflection, accessibility, and recognition of conversational AI’s limited empathy and dependency risks. The findings position conversational AI as a perceived supplementary outlet rather than evidence of improved mental health or a substitute for authentic relationships and professional care.