The Effect Of A Social Robot's Autonomy And Group Orientation On Human Decision-making

This study analyzed the influence of a social robot on human decision-making. This study tried to explore how the following factors influence human decision-making in response to a social robot: 1) the level of autonomy (high vs. low) of a robot; 2) a robot's group orientation (ingroup vs. outgroup); 3) individual's self-construal; 4) individual's negative attitudes toward robots.With this objective, a human-robot interaction experiment was designed, combing both the quantitative questionnaire measurements as well as the qualitative observational data. In the experiment, participants completed a collaborative task together with a social robot and the robot provided suggestions to the participants. The level of autonomy and group orientation of the robot were the between-subject independent variables; gender, self-construal, negative attitudes toward robot, accuracy of the initial decisions were the covariates; the robot's influence on decision-making, faith, attachment, credibility, social presence, mental workload and interactive behaviors were taken as dependent variables.There were 48 participants attending the experiment. The result indicated that participants were more likely to accept recommendations from a robot with high level of autonomy, and they showed higher faith to it and had longer gazing time to it. Participants had lower mental workload when the robot was an ingroup member than when it was an outgroup member. The high level of autonomy increased participants' acceptance of a robot's recommendations more when the robot was an ingroup member than when it was an outgroup member. The male participants considered the robot more trustworthy than female participants did, but they shared less benefit to the robot compared with the female counterparts; the negative attitudes toward the social impact of robots predicted participants' mental workload in the interaction with the robot.Based on the study, the following suggestions were provided for social robot designers: 1) when a robot is highly reliable and high influence on human decision-making is desired, it could be designed more autonomous. 2) A robot could be characterized as an ingroup member to lower the mental workload. 3) When the robot is an ingroup member, the level of autonomy of the robot should be carefully designed to adjust users' acceptance of its recommendations. 4) Individual's attitudes toward robots influence his/her interaction with robot and designers should consider this factor.