Using emojis to uplift emotion: Effects on mental health chatbots’ performance

As AI-based chatbots increasingly expand access to mental health support, understanding how text-based systems convey emotional attentiveness without face-to-face nonverbal cues is important. Emojis may serve as lightweight paralinguistic cues, yet their effects in mental health support remain unclear. Drawing on the Stereotype Content Model and psychological ownership theory, this study examines associations between emoji use and users’ emotional well-being, satisfaction, and continuance usage intention. Two scenario-based online experiments were conducted in China using random sampling through Credamo, yielding 459 valid responses (59 in the pretest and 400 in the main study). Data were analyzed using IBM SPSS Statistics 26 and the PROCESS macro. Across the experiments, emoji use was associated with more favorable emotional well-being, satisfaction, and continuance usage intention. Mediation analyses supported an indirect pathway through perceived warmth and psychological ownership, whereas the corresponding competence-based pathway was not supported. AI literacy strengthened the perceived warmth–psychological ownership association but, contrary to predictions, weakened the perceived competence–psychological ownership association. Moderated mediation analyses provided preliminary support for AI literacy strengthening the warmth-based indirect effects of emoji use on emotional well-being and continuance usage intention. The corresponding effect on satisfaction was not supported. These findings extend research on AI-mediated mental health support by identifying a warmth–psychological ownership pathway and the differentiated role of AI literacy. The study also offers preliminary, context-bounded guidance for the strategic and context-sensitive use of emojis in supportive, non-clinical chatbot interactions.