Towards a Model for Building Trust and Acceptance of Artificial Intelligence Aided Medical Assessment Systems
(A2020-64418)

AUTHORS

Lennart Seitz, Leuphana University Lüneburg; Sigrid Bekmeier-Feuerhahn, Leuphana University Lueneburg; Florian Bontrup, Leuphana University Lueneburg; Justus Wildt, Docyet; Krutika Gohil, Docyet

KEYWORDS

chatbots; healthcare; trust

ABSTRACT

This study aims to identify determinants for the emergence of trust in AI-based medical assessment systems consisting of chatbots and telemedicine. Existing studies have been failing to create a holistic understanding due to focusing on single trust antecedents. Our study closes this research gap by conducting semi-structured interviews and standardized questionnaires to identify relevant variables and their relationship to each other. Participants (n = 40) take part in a laboratory experiment interacting with a chatbot (vs. chatbot + human agent) for initial medical assessment. The first results indicate the importance of the chatbot’s purpose and the transparency of underlying data base. Furthermore, communication patterns conveying uncertainty reduction are found to be more important than chatbot’s social skills. The additional human expert complements the chatbot due to the possibility of more specific and detailed questioning and patients’ wish of having a responsible person.