PROCEEDINGS OF THE ANNUAL CONFERENCE OF JSSD
THE 71st ANNUAL CONFERENCE OF JSSD
Session ID : D8-04
Conference information

Using generative AI to Design Approachable Avatars
*Teru KawakitaNarumi Kawakita
Author information
CONFERENCE PROCEEDINGS FREE ACCESS

Details
Abstract

Using generative AI, we explored the design of counselor avatars (AI agents) that facilitate easy sharing of minor concerns. In this study, we created male/female avatars with different appearances: anime styles, 2D styles, 3D styles, semi-realistic styles, and photorealistic styles. This study was conducted with 18 female students to determine which avatar they found easiest to talk to. The results showed that female avatars in anime style, 2D style, and photorealistic style received high ease-of-communication ratings. Conversely, some male avatars tended to receive lower ratings.

Content from these authors
© 2024 JAPANESE SOCIETY FOR THE SCIENCE OF DESIGN
Previous article Next article
feedback
Top