Conversational agents such as chatbots have emerged as a useful resource to access real-time health information online. Perceptions of trust and credibility among chatbots have been attributed to the anthropomorphism and humanness of the chatbot design, with gender and race influencing their reception. Few existing studies have looked specifically at the diversity of chatbot avatar design related to both race, age, and gender, which may have particular significance for racially minoritized users like Black older adults. In this paper, we explored perceptions of chatbots with varying identities for health information seeking in a diary and interview study with 30 Black older adults. Our findings suggest that while racial and age likeness influence feelings of trust and comfort with chatbots, constructs such as professionalism and likeability and overall familiarity also influence reception. Based on these findings, we provide implications for designing text-based chatbots that consider Black older adults.
https://doi.org/10.1145/3544548.3580719
The ACM CHI Conference on Human Factors in Computing Systems (https://chi2023.acm.org/)