The Influence of Perceived Fairness on Trust in Human-Computer Interaction

Area of Science:

• Psychology
• Human-Computer Interaction
• Artificial Intelligence Ethics

Background:

• Fairness is a core social interaction principle impacting decisions.
• Human-computer interaction is a growing area of social engagement.
• Understanding trust in AI is crucial as artificial intelligence (AI) becomes more prevalent.

Purpose of the Study:

• To investigate fairness perceptions in human-human versus human-AI interactions.
• To analyze how fairness influences trust decisions in both contexts.
• To compare trust investment rates and amounts based on proposer and trustee identity (AI/human) and offer fairness.

Main Methods:

• A mixed experimental design with 2 (proposer identity: AI/human) × 2 (offer: fair/unfair) × 2 (trustee identity: AI/human).
• Utilized the Ultimatum Game and Trust Game paradigms.
• 128 university students participated in the study.

Main Results:

• Fair offers led to higher trust investment rates and amounts than unfair offers.
• When offers were unfair, participants showed greater investment willingness towards AI proposers than human proposers.
• Under fair conditions, participants were more risk-averse with human trustees than AI trustees.

Conclusions:

• Fairness perceptions in human-computer interactions have a stronger impact on trust decisions than in human-human interactions.
• AI's role in fairness and trust dynamics differs from human-human interactions.
• Findings highlight the importance of fairness in building trust in AI systems.