Uncanny Valley Effect is Amplified with Multimodal Stimuli and Vary Across Ages
“Uncanny Valley Hypothesis” suggests that humanoid objects that materialize human beings virtually but not entirely realistically may elicit uncanny feelings of eeriness and revulsion in observers. While the uncanny valley (UV) has been largely investigated with a focus on the visual aspects of the robot-like designs with young adults, the auditory components that may contribute to this effect and how visual and auditory factors jointly play a role in uncanny reports across different generations has not been examined. In the present study, we investigated how multimodal stimuli and the congruence of visual and auditory aspects of the stimuli contribute to the uncanniness perception and differ from the audio and visual components across generations. Young and old adults rated animations that were presented in audio-visual, audio-only and visual-only modalities in terms of uncanniness. The visual and auditory aspects of the stimuli had four levels of naturalness: robot (unrealistic), semi-robot (semi-realistic), human-like (realistic) and human (real). Our results show that audio-visual stimuli have an amplified effect on UV scores than only auditory and only visual stimuli. In addition, multimodal stimuli that have incongruent audio and visual components elicited significantly higher uncanny scores than stimuli that have congruent components. However, the difference between congruent and incongruent stimuli were more pronounced in the younger group compared to the older group. We also found that younger generations are more sensitive to naturalness layers of audio-visual stimuli than older generations. In conclusion, uncanny valley effect is modulated by stimulus modality, congruence of visual and auditory modalities, naturalness as well as age.