We propose a “role-play AI” that can respond to inquiries, receptions, and guides at city offices. Conventional AI training requires a lot of accurate training data, and data collection has been an extremely costly and difficult task. In this study, we solve this problem by utilizing community cooperation. By making the data collection work a community cooperation activity, we collect accurate training data at a low cost. By connecting “people who live in the area” and “people who interested in the area,” we have collected very high-quality training data, and we have made possible the training of role-play AI that is closely linked to the community. By using this technology, we provide “role-play AI” to learn according to local demand. In the future, we aim to realize AI technology that can be used in a wide range of situations, not only by local governments.
/ Interaction Research Group, Innovative Communication Laboratory