Do we want AI judges? The acceptance of AI judges’ judicial decision-making on moral foundations

This study explored the acceptance of artificial intelligence-based judicial decision-making (AI-JDM) as compared to human judges, focusing on the moral foundations of the cases involved using within-subject experiments. The study found a general aversion toward AI-JDM regarding perceived risk, permissibility, and social approval. However, when cases are rooted in the moral foundation of fairness, AI-JDM receives slightly higher social approval, though the effect size remains small. The study also found that demographic factors like racial/ethnic status and age significantly affect these attitudes. Especially, racial/ethnic minorities showed more social approval in AI-JDM than the racial/ethnic majority (i.e., non-Hispanic whites). These findings show the potential role of machine heuristics in the acceptance of AI-JDM and have practical implications for the implementation of AI in judicial systems, particularly suggesting a cautious approach and underscoring the need to inform the public, especially minority groups, about the limitations and potential biases of AI-JDM.