의인화된 에이전트 중심의 인공지능 홈트레이닝 서비스 개선

Abstract

With the increasing influence of untact worldwide, a variety of home training-related products and services are being released. Among them, a service called Artificial Intelligence Home Training is providing new experiences to users focusing on artificial intelligence home trainers. Studies have shown that these anthropomorphic agents provide users with positive use experiences, creating psychological bonds. Thus, this study classified offline personal training and online content-based home training as traditional training types, and interviewed to derive eight strengths of real trainers, which can be called mental models for users. Since then, 11 personalization-related requirements have been collected for groups with experience in artificial intelligence home training services and responded to the strengths of real trainers. Then, we came up with an idea to increase the level of agent personification. This was organized into six stages of use to produce prototypes and to determine their usefulness, satisfaction, and sustainability for verification of the design. As a result, the respondents showed relatively high usefulness, satisfaction, and sustainability in the "set-up for users" and "talking with artificial intelligence trainers" stages, where artificial intelligence trainers and interactions are most active.