The Effect Of Uncanny Valley Effect On Robot Moral Judgment

Artificial intelligence and robotics are rapidly evolving.Autonomous robots can slowly make choices with moral consequences,and humans are increasingly affected by them.At the same time,research on robots has shown that when robots resemble humans to a certain degree,people’s reactions to them become negative and repulsive-a phenomenon known as the uncanny valley effect.However,little is known about whether the moral judgments of robots are affected by the uncanny valley effect.Therefore,we investigated it through three experiments.Experiment 1 uses a questionnaire to investigate whether people’s moral judgments about robots’ harming behaviors are affected by the uncanny valley effect.Experiment 2 used a questionnaire survey to investigate whether people’s moral judgments about making decisions about robotic agents in moral dilemmas,which were divided into instrumental and contingent dilemmas,would be affected by the uncanny valley effect.Based on the results of Experiment 2,Experiment 3 explored the mechanisms behind moral judgments about terroristic robots based on dual processing theory,and added female agents to the types of agents.The results of Experiment 1 showed that the main effect of victim type was significant in scoring the severity of the agent’s behavior,and harming humans was perceived as more serious.The interaction between victim type and agent type was significant.The main effect of victim type was significant in scoring the severity of the agent’s behavior,and harming humans was considered more severe.The victim type interacted significantly with the agent type.In the condition that the victim was a robot,sentencing was significantly higher for the terror type robot than for the other agents.The uncanny valley effect influenced subjects’ judgments.Mental perceptions of agents did not differ significantly in the subjectivity latitude,and robots scored significantly lower than humans in the perception latitude.Higher scores were associated with a higher perceived ability of the agent to feel emotions,implying that subjects perceived the robot to be as autonomous as humans in making decisions,but not as capable of feeling as humans.Analysis of the individual dark triad levels showed that the higher the individual’s psychopathy score was when the human was the victim,the more indifferent the individual was to harming humans.This is consistent with the results of previous studies.In addition to this,it was also found that subjects with high levels of narcissism showed very low tolerance for behaviors that harm humans,and subjects who scored higher in narcissistic latitude would score and sentence higher and heavier the severity of the agent’s behavior in the condition that the victim was a human.The results of Experiment 2 showed that the uncanny valley effect did not have an effect on subjects’ judgments,differing significantly only in the classic footbridge dilemma,but we found significant differences by gender in moral judgments about different agents.Women preferred utilitarianism and men preferred the moralistic choice,and this difference was not affected by dilemma type.The results of Experiment 3 showed that the type of dilemma had an effect on subjects’ reaction time and decision making,but there was a significant difference only when the human agent group reacted.The uncanny valley effect did not cause subjects to make more utilitarian judgments about the robot and to differ significantly from other agents in response time,as we expected,but consistent with robots with a high degree of anthropomorphism in appearance,subjects’ judgments about robots whose appearance elicited the uncanny valley effect were similar to those of human agents in response time and decision preferences,and differed from those of robots that did not resemble humans.This suggests that although the uncanny valley effect elicits creepy and disgust,it still makes the same judgments as humans because of its more human-like appearance.Based on the results of this study,the following conclusions can be drawn: first,the type of victim affects people’s moral judgment of the agent,and the uncanny valley effect affects people’s judgment of the robot agent only when the victim is a robot;second,moral judgment of the terroristic robot in moral dilemmas is similar to that of humans;third,there are gender differences in moral decision making about the robot,with men preferring deontology;fourth,the gender of the agent in the dilemma does not affect people’s moral decision making.