Dynamic Fairness Perceptions in Human-Robot Interaction

TL;DR

The results show that fairness judgments can shift based on the timing of unfair robot actions, and explore using perceptions of three key factors proposed by a Fairness Theory from Organizational Justice to predict momentary perceptions of fairness in this study.

Abstract

People deeply care about how fairly they are treated by robots. The established paradigm for probing fairness in Human-Robot Interaction (HRI) involves measuring the perception of the fairness of a robot at the conclusion of an interaction. However, such an approach is limited as interactions vary over time, potentially causing changes in fairness perceptions as well. To validate this idea, we conducted a 2x2 user study with a mixed design (N=40) where we investigated two factors: the timing of unfair robot actions (early or late in an interaction) and the beneficiary of those actions (either another robot or the participant). Our results show that fairness judgments are not static. They can shift based on the timing of unfair robot actions. Further, we explored using perceptions of three key factors (reduced welfare, conduct, and moral transgression) proposed by a Fairness Theory from Organizational Justice to predict momentary perceptions of fairness in our study. Interestingly, we found that the reduced welfare and moral transgression factors were better predictors than all factors together. Our findings reinforce the idea that unfair robot behavior can shape perceptions of group dynamics and trust towards a robot and pave the path to future research directions on moment-to-moment fairness perceptions

Figures (4)

• Figure 1: The study took place in a laboratory (a). The participant played a multi-player Space Invaders game game with two robots (b). Nao's spaceship (white) supported players by eliminating enemies on their side of the game screen. The lead marker moved to the side of the player with the highest score. Enemies in the game reappeared after being eliminated. The Nao robot moved its head to face the player whose spaceship it supported and made comments to convey that it was trying to help them (c). The Shutter robot was expressive through four different behaviors controlled via a behavior tree control architecture (d).
• Figure 2: Timeline of a study session (a). The study began with an introduction to the robots and two practice games to learn about Space Invaders. Then, during Game Round 1, participants experienced a 2 minute game where Nao equally supported each player by switching its support when it eliminated 10 enemies on one side. In game rounds 2 and 3, the participants experienced Nao providing unequal support. For the first or the second half of these games, Nao focused on helping only Shutter (b) or the human participant (c). We measured impressions of the experience for each game round. Early and late unfair support (in the second and third game rounds) was counterbalanced across participants to reduce potential ordering effects.
• Figure 3: Left: Average momentary fairness ratings (error bars are standard error). Right: Overall fairness ratings. The boxes represent the interquartile ranges of ratings. The dots outside the boxes denote outliers.
• Figure 4: Average percentages of "Not Applicable"(NA) in Trust Scales. The boxes represent the interquartile ranges of ratings. The dots outside the boxes denote outliers.