Social, but still uncanny

Abstract

The Uncanny Valley hypothesis proposes that as robots become more human-like, they are initially liked better but then elicit a feeling of eeriness, peaking just before achieving full human resemblance. It remains unclear whether context can modify this effect. In an online experiment, participants were primed with a vignette about either robots as social companions (social context priming) or a neutral topic, and then rated images of robots on human-likeness, likability, trust, and creepiness. We found a negative linear relationship between a robot’s human-likeness and its likability and trustworthiness and a positive lin-ear relationship between a robot’s human-likeness and creepiness. Social context priming improved overall likability and trust of robots but did not modulate the Uncanny Valley effect. This indicates that, while presenting robots in a social context can improve their acceptance, this does not change our inherent discomfort with increasing human-like robots.