Reducing consistency in human realism increases the uncanny valley effect; increasing category uncertainty does not

Area of Science:

• Psychology
• Human-Computer Interaction
• Robotics

Background:

• The uncanny valley describes the eerie feeling evoked by near-human replicas.
• Masahiro Mori (1970) linked this to inconsistent realism between human and non-human features.
• Ernst Jentsch (1906) proposed ambiguity and category uncertainty as drivers of unease.

Purpose of the Study:

• To investigate if reduced realism consistency in visual features amplifies the uncanny valley effect.
• To test competing theories on the causes of the uncanny valley phenomenon.

Main Methods:

• Conducted three experimental rounds with 548 participants.
• Participants categorized and rated stimuli (humans, animals, objects) varying in realism.
• Manipulated realism consistency in specific feature sets (e.g., eyes-mouth, skin-nose).

Main Results:

• Reduced realism consistency significantly increased eeriness and coldness perceptions for humans and animals.
• Objects did not exhibit a similar increase in negative emotional response.
• High category uncertainty did not correlate with heightened uncanny valley perceptions, refuting Jentsch's theory.

Conclusions:

• Visual realism inconsistency is a key factor driving the uncanny valley effect in human and animal replicas.
• The findings support Mori's theory over Jentsch's regarding the primary cause of the uncanny valley.
• Understanding these perceptual mechanisms is crucial for designing more acceptable artificial entities.