AI-Mediated Dispute Resolution: Trust, Fairness, and the Limits of Automation

Abstract: As artificial intelligence (AI) becomes increasingly embedded in dispute resolution, questions arise about whether AI can meaningfully substitute for human mediators. Although AI-mediated tools promise efficiency and scalability, mediation effectiveness depends on trust, procedural fairness, and emotional dynamics. Across three studies, we (will) examine how mediator identity (AI vs. human) and dispute context shape disclosure, perceptions of fairness, trust in the mediator, and agreement outcomes. Study 1 shows that participants prefer AI mediators for transactional disputes but favor human mediators for emotionally charged conflicts. Paradoxically, participants report greater emotional openness with AI despite lower perceived benevolence. Preliminary evidence in a pilot of Study 2 provides causal evidence that AI increases disclosure, whereas human mediation enhances perceived voice and relational satisfaction. A proposed Study 3 tests dispute context as a boundary condition using an operational AI mediation platform.

Keywords: Artificial intelligence; mediation; conflict; trust; fairness