It’s all about respect: How explanations impact trust and procedural justice in AI-assisted judicial decision-making

Abstract

As artificial intelligence (AI)-assisted judicial decision-making (AIJDM) becomes more common globally, understanding what affects court users’ trust in processes where this technology is used becomes crucial for maintaining judicial legitimacy. This study investigates how different types of explanations about AI for assisting judicial decision-making affect trust through procedural justice mechanisms. Using a 3 × 2 between-subjects experiment ( n  = 859), we measured trust in judicial decision-making processes where AI was used across six conditions, varying the explanation type about the AI – process explanations (how the AI system works), outcome explanations (how it reached a specific recommendation on a decision) or no explanation – and the decision context (liability vs. quantum). Mediation analysis showed that providing explanations (regardless of type) substantially increases both cognitive and affective trust compared to no explanation, in both liability and quantum decision-making contexts. Critically, procedural justice perceptions strongly mediated this effect, with respect emerging as the dominant mediator (accounting for 61–72% of indirect effects), followed by neutrality, while voice contributed negligibly. This mediation was particularly strong for affective trust (89.6% of total effect) compared to cognitive trust (69.8%). These findings extend procedural justice theory to AI-assisted judicial contexts and suggest that communicating explanations about AIJDM is a vital mechanism for signaling respect to participants in a court process and maintaining judicial legitimacy in increasingly AI-assisted judicial processes.