The roles of cognitive dissonance and normative reasoning in attributions of minds to robots

Abstract

As a wide variety of intelligent technologies become part of everyday life, researchers have explored how people conceptualize agents that in some ways act and think like living things but are clearly machines. Much of this work draws upon the idea that people readily default to generalizing human-like properties to such agents, and only pare back on these generalizations with added thought. However, recent findings have also documented that people are sometimes initially hesitant to attribute minds to a machine but are more willing to do so with additional thought. In the current experiments, we hypothesized that these attribution-increasing reconsiderations could be spurred by situation-induced cognitive dissonance. In two experiments, participants completed a belief activation exercise designed to induce cognitive dissonance (writing arguments for or against prominent beliefs), viewed a video of an ambiguously intentional robot, and completed measures of cognitive dissonance. In both experiments, cognitive dissonance was associated with increased attributions of mind to the robot. Our findings provide evidence that people sometimes increase their attributions of minds when experiencing cognitive conflict, but also that activation of change-inducing concepts may impact attributions of a mind without producing conscious cognitive conflict in participants.