The unconscious effect of physical beauty in human social interactions

Supervisors: 

Pablo Arias Sarah, MVLS/School of Psychology and Neuroscience 

Professor Alessandro Vinciarelli, CoSE/ School of Computing Science 

Mathieu Chollet, CoSE/ School of Computing Science

 

PHD Project Summary:

Transformation filters have invaded our digital lives. It is now possible to clone a person’s voice, manipulate their facial movements, change their vocal emotions, and more. These transformations are progressively being integrated in large scale consumer applications (e.g., Instagram, TikTok, Snapchat) and remote communication software (e.g. Zoom), creating a context where AI-based transformation filters are available to billions of (the youngest) individuals worldwide. However, the manipulative potential of these transformations is still unkown. Investigating that question is the aim of the current studentship.

To do so, we recently developed a video-conference experimental platform called DuckSoup, which enables researchers to manipulate the social signals produced by participants in real-time, and without their awareness (e.g. increasing or decreasing their smiles). The goal of this scholarship is to develop a face transformation algorithm able to manipulate facial beauty in real time, and to examine whether and how it influences human behavior during social interactions (e.g. job interviews, negotiation).

Our aim is to uncover how facial beauty unconsciously impact (1) human social decisions and (2) the multimodal behavior of participants, which we measure through the analysis of voice, face, semantic and physiological recording