[3I6-GS-11] AI and Society:

[3I6-GS-11-02] Decision-Making Support Through Multi-Robot Interaction with Diverse Perspectives

〇Satoshi Nakagawa¹, Misato Nihei¹ (1. The University of Tokyo)

Keywords:Human-robot Interaction, Well-being, Decision Making, Conformity Pressure, Assistive Technology

In online communication, conformity pressure significantly influences individual decision-making, often leading to biased judgments and reinforcing social polarization. This study investigates whether interactions with conversational robots of differing personalities can mitigate such effects and promote diverse perspectives. We conducted an experiment where participants observed a biased group chat and subsequently engaged in dialogues with two types of robots: a conformity-driven robot and a critical-thinking robot that challenges biases. Results revealed that participants initially exhibited conformity pressure, but exposure to multiple robot perspectives encouraged participants to express independent opinions. In contrast, the conformity-driven robot reinforced existing biases. These findings suggest that exposure to multiple perspectives via conversational AI can reduce cognitive biases and improve decision-making flexibility. Our research contributes to AI-assisted dialogue systems that foster well-balanced discussions, ultimately aiding in the realization of a well-being centered society.