Caring for Premenstrual Discomfort with “Companions”: Novel “Peer” Support with Multiple Chatbots

A research group led by Associate Professor Koji Yatani at the University of Tokyo’s Graduate School of Engineering has shown that a “group-based conversational system,” in which multiple chatbots have conversations with a user can effectively alleviate the physical and emotional discomfort associated with premenstrual syndrome (PMS).

During a two-month study involving 63 participants, the researchers found that the system fostered a reassuring sense of camaraderie and encouraged users to reflect on their own symptoms through the diverse perspectives shared by chatbots acting as peer participants. Compared with conventional one-on-one chat systems, the group format increased the volume of user utterances, boosted engagement, and encouraged exploration of a wider range of coping strategies. The division of roles among the chatbots also enabled the delivery of both concrete suggestions and empathetic responses across a broader spectrum of needs.

Because the system allows users to engage in group-style conversations even when going out or seeking in-person counseling is difficult, it has the potential to reduce feelings of isolation and the reluctance to seek help caused by PMS, thereby creating a more supportive environment for many women.

Figure. A conceptual comparison of an 1on1 and group style conversation environment examined in our work. Our system enables a group conversation by bringing multiple chatbots which play different roles. In this manner, our system offers group interaction experience while maintaining psychological and privacy safety as all the other entities are bots.

Papers
Conference: 2025 CHI Conference on Human Factors in Computing Systems (ACM CHI 2025)
Title: Beyond the Dialogue: Multi-chatbot Group Motivational Interviewing for Premenstrual Syndrome (PMS) Management
Authors: Shixian Geng*, Remi Inayoshi, Chi-Lan Yang, Zefan Sramek, Yuya Umeda, Chiaki Kasahara, Arissa J. Sato, Simo Hosio, and Koji Yatani*
DOI: 10.1145/3706598.3713918