Beyond Attention: Investigating the Threshold Where Objective Robot Exclusion Becomes Subjective

TL;DR

This study investigates social exclusion in robot-led group interviews using Ameca, distinguishing objective exclusion (robot attention distribution) from subjective exclusion (perceived exclusion) within the Temporal Need-Threat Model. It combines eye-gaze–driven attention metrics, subjective reports, mood and need fulfillment measures, and a threshold analysis to identify when objective exclusion becomes subjectively salient, finding a critical point at $0.894$ for objective exclusion. Mediational analyses show subjective exclusion mediates effects on need fulfillment (but not mood), while objective exclusion directly predicts subjective exclusion; standing position emerges as the primary risk factor for both objective and subjective exclusion, with demographic factors showing negligible effects. The results highlight the importance of accounting for subjective perception and spatial positioning in designing fair and inclusive robot-assisted hiring and group-interview interactions, and they propose design-focused mitigations such as a rectified eyegaze control algorithm to reduce spatial bias.

Abstract

As robots become increasingly involved in decision-making processes (e.g., personnel selection), concerns about fairness and social inclusion arise. This study examines social exclusion in robot-led group interviews by robot Ameca, exploring the relationship between objective exclusion (robot's attention allocation), subjective exclusion (perceived exclusion), mood change, and need fulfillment. In a controlled lab study (N = 35), higher objective exclusion significantly predicted subjective exclusion. In turn, subjective exclusion negatively impacted mood and need fulfillment but only mediated the relationship between objective exclusion and need fulfillment. A piecewise regression analysis identified a critical threshold at which objective exclusion begins to be perceived as subjective exclusion. Additionally, the standing position was the primary predictor of exclusion, whereas demographic factors (e.g., gender, height) had no significant effect. These findings underscore the need to consider both objective and subjective exclusion in human-robot interactions and have implications for fairness in robot-assisted hiring processes.

Figures (5)

• Figure 1: Temporal Need-Threat Model (TNTM) (adapted from williams2009ostracism)
• Figure 2: Social humanoid robot Ameca from Engineered Arts (left) used in the exclusion study; mock interview flow with the robot (right).
• Figure 3: Higher correlation indicating potential mediation effects.
• Figure 4: Results of Piecewise Regression.
• Figure 5: Standing Position as a Risk Factor for Robot-Induced Exclusion.