Research on Social Familiarity and Self in the Uncanny Valley: A Multidimensional Exploration (Master Thesis)

Author & Lead Researcher | Advisor: Prof. Guomei Zhou
Duration: December 2024 – Present

This ongoing master’s thesis explores the uncanny valley effect through the lenses of social cognition and self-concept, with an emphasis on the interplay between human-likeness, social familiarity, and self-referential cues. The project also involves the creation and validation of a novel psychological scale targeting human-robot interactions.

Current Progess

• Behavioral Experiments:
  Designed and conducted a series of experiments manipulating human-likeness, social familiarity, and self-referential stimuli to test major uncanny valley hypotheses and examine affective and cognitive responses

• Data Modeling & Analysis:
  Applied advanced statistical modeling techniques, including LASSO regression and mixed-effects models, to analyze experimental data. Developed a multidimensional conceptual framework to interpret behavioral patterns, with extensive use of data visualization

• Scale Development & Validation:
  Conducted an extensive literature review to refine the theoretical basis and identify key constructs for a new scale measuring human threat perception and defensive mechanisms toward robots. Established the scale’s reliability using exploratory (EFA) and confirmatory (CFA) factor analyses for validity analysis is planned

Next Steps in My Research

• Employ machine learning to model and predict individual differences in uncanny valley responses
• Systematically examine how the uncanny valley manifests in AI-generated faces, including temporal and contextual influences
• Utilize VR environments to study naturalistic, real-time human–AI social exchanges