A Robot’s Justifications, but not Explanations, Mitigate People’s Moral Criticism and Preserve Their Trust

Abstract

In two experiments we investigate people’s moral criticism (blame) of social robots that violate norms, as well as the impact of the robot’s justification of its norm violation on people’s blame and trust. We found consistent evidence of a “disobedience” effect—greater blame for robot decisions that deviated from what participants suggested the robot should do. Further, we found that justifications differentially affected blame and trust. They mitigated blame only for certain decisions but raised trust regardless of the decision. Justifications may signal that robots are morally trustworthy even if they made a specific decision that deserved blame.