More than appearance: the uncanny valley effect changes with a robot’s mental capacity

Robots are being used to socially interact with humans. To enhance the quality of human-robot interaction, engineers aim to build robots with both a humanlike appearance and high mental capacity, but there is a lack of empirical evidence regarding how these two characteristics jointly affect people’s emotional response to robots. The current two experiments (each N = 80) presented robots with either a mechanical or humanlike appearance, with mental capacities operationalized as low or high, and with either self-oriented mentalization to mainly concentrate on the robot itself or other-oriented mentalization to read others’ minds. It was found that when the robots had a humanlike appearance, they were more dislikeable than when they had a mechanical appearance, replicating the uncanny valley effect for appearance. Importantly, given a humanlike appearance, robots with high mental ability elicited stronger dislike than those with low mental ability, showing an uncanny valley effect for mind, but this difference was absent for robots with a mechanical appearance. In addition, this effect was limited to robots with self-oriented mentalization ability and did not extend to robots with other-oriented mentalization ability. Hence, the exterior appearance and interior mental capacity of robots interact to influence people’s emotional reaction to them, and the uncanny valley as it pertains to the mind depends on the robot’s appearance in addition to its mental ability. This implies that social robots with humanlike appearances should be designed with obvious other-directed social abilities to make them more likeable.