Beyond Gender Stereotypes: Evaluating Trust and Value in Well-Being Chatbot Interactions. Extended Abstract

Healthcare chatbots improve mental health support through 24/7 availability, reduced stigma and cost, and proven effectiveness, yet adoption is limited by perceptions of low humanity. Research based on CASA paradigm and Stereotype Content Model indicates gender cues may affect perceived warmth and competence, driving interaction value, trust, and use. However, evidence for counter-stereotypical effects in healthcare remains limited. This research investigated whether chatbot gender influences perceived warmth, competence, trust, value, attitudes, and usage intention in well-being contexts, examining warmth and competence's roles in trust and value generation. Two experiments manipulated chatbot gender through names and avatars while keeping other features constant. Study 1 used scenarios with 297 Prolific participants evaluating chatbot screenshots. Study 2 replicated the test with 474 Prolific participants interacting with a well-being chatbot providing exercise recommendations. Both studies used validated 5-point Likert measures, analyzing factor scores through ANOVAs and regression models with successful manipulation checks. Contrary to gender-stereotype expectations, male- and female-gendered well-being chatbots showed no significant differences in perceived warmth, competence, attitudes, or usage intentions. Warmth and competence increased trust, which mediated attitudes and intentions. Both dimensions enhanced functional and emotional value, with emotional value showing stronger influence on user attitudes and perceived well-being. The findings suggest designers should prioritize trust-building through warmth and competence and emotional resonance rather than gendered humanization cues. Future work should examine richer gender manipulations and user-chatbot gender congruence