Abstract
Although many chatbot applications have been proposed to support mental health care, the effective interaction design for chatbots to assist positive thinking for problem-solving without relying on the content of individual consultation is not clear. In this paper, we developed a chatbot application that can perform conversation tasks based on the Miracle Question, a method of the solution-focused approach. Then we validated that solution-focused interactive task on chatbot helps improve well-being index and active behavior. As a result, the results did not confirm that the proposed conversation task significantly improved the scores on each scale. However, when compared to existing tasks for behavior activation support, the proposed task tended to have a positive effect on improving well-being scores. Our finding has implications for the methodology and difficulties in supporting personal problem-solving based on minimal design.
