Decision-Making Model for Robots that Consider Group Norms and Interests

Abstract

In this study, we propose a decision-making model that enables a robot to behave socially in groups of people with common interests. Several attempts have been made to develop robots suitable for humans. Robots should act socially in a human-like social manner to help humans live comfortably. However, if robots are completely subservient to humans, it may not have only a positive impact on the humans’ social life among their human communities. It is crucial to create robots that are not extremely comfortable for humans as partners. Thus, we propose and evaluate a decision-making model for a robot that can act selfishly and altruistically in a human–robot scenario, while considering mutual interests. This is expected to contribute to the robot’s decision-making design in a society where humans and robots coexist in the future. Further, we investigated how a virtual robot equipped with the proposed interest-aware model behaves in interactions with humans in the proposed group-type ultimatum game. The results of the investigation indicate that the robot equipped with our proposed model can combine group and self-oriented behavior in an experimental scenario.