Abstract: The perception of agency in robots and AI characters has become increasingly important as different agents increase their capabilities. Experiment 1 took an existing measure of perceived agency and created a reduced version by using existing Rasch item reduction measures; Eight and five item scales were created. Experiment 2 showed that all three scales (PA, PA8, PA5) were able to capture differences in perceived agency between a cheating robot (higher PA) and a non-cheating robot (lower PA). Experiment 3 showed that all three scales were able to show the predicted positive relationship between perceived agency and perceived moral agency. All three scales also showed high internal validity. Suggestions for the usage of the scales was also discussed.
Loading