Prediction of Trustworthiness from Human Demographic & Behavioral Information of An Audio-Visual Corpus

The establishment of relationships relies fundamentally on trust. Investigating the attributes contributing to trustworthiness holds significant merit for enhancing decision-making processes across social, political, business, education, and healthcare domains. A critical challenge lies in discerning how individuals make trust-related decisions and the specific cues influencing such judgments. This research endeavors to examine the interplay between human demographic and behavioral information and their correlation with trustworthiness. A diverse set of participants engaged in a survey, evaluating short acted emotional videos based on attributes including authenticity, comfort, eloquence, expressiveness, kindness, physical attractiveness, trustworthiness (TW), and voice attractiveness. Through Spearman rank correlation analysis, authenticity (ρ = 0.68) and kindness (ρ = 0.66) emerge as the primary attributes strongly associated with TW. Conversely, the demographic background of annotators demonstrates a weak relationship with TW.