Contextual justice: citizens’ dialogues on AI judges in Japan

Artificial intelligence (AI) is increasingly discussed as a tool for judicial decision-making, yet there is limited empirical evidence on how citizens in Japan evaluate the acceptability of AI judges across concrete cases. This paper examines how participants in a Japan-based philosophy café evaluated AI judges across abstract attitude questions and three case scenarios involving hunger-driven theft, wrongful accusation, and heinous crime. Sixty participants selected an AI judge, a human judge, or both, and explained their choices in writing. A majority anticipated the future use of AI judges and expressed support for such use in response to the abstract support question. However, preferences shifted across scenarios: participants favored human judges in the hunger-theft case, whereas responses were more mixed in the wrongful-accusation and heinous-crime cases. Also, the proportion choosing “either” decreased as case severity increased. Thematic analysis identified three recurring clusters of justification: instrumental considerations; procedural justice and interpersonal qualities; and responsibility, control, and trust in human–AI collaboration. Empathy-related concerns were especially prominent in justifications for human judges, whereas fairness- and procedure-related considerations were linked to openness to AI in some contexts. These findings highlight the context sensitivity of public evaluations of AI judges and the value of examining scenario-based, citizen-centered judgments alongside abstract attitudes. They also have implications for AI governance in legal systems.