Abstract
To use AI systems that are trustworthy, it is necessary to consider not only AI technologies, but also a model that takes into account factors such as guidelines, assurance through audits and standards and user interface design. In this paper, we conducted a questionnaire survey focusing on (1) AI intervention, (2) data management, and (3) purpose of use. The survey was conducted on a case study of an AI service for dietary habit improvement recommendations among Japanese people. The results suggest that how the form of communication between humans and AI is designed may affect whether users trust and use AI.