Following AI or Humans? The Effect of Perceived Similarity on Judges’ Advice Utilization in Ethical Decisions

As AI systems increasingly integrate into decision-making processes across various sectors, their advisory role in ethically charged decisions remains an understudied topic. This research in progress investigates the dynamics of Human-AI hybrid Judge-Advisor Systems (JAS), focusing on the influence of perceived similarity between judges and advisors on decision outcomes. Specifically, it examines how the type of advisor (AI vs. human) and the content of the recommendation (rule-based vs. empathy-based) affect judges’ likelihood of accepting or overriding recommendations. Preliminary findings suggest that judges are more inclined to override AI’s rule-based advice compared to human advice, while empathy-based recommendations yield no significant difference. These insights are explained through perceived similarity, highlighting the complex interplay between AI’s perceived capabilities and human ethical judgment. The study has the potential to contribute to the understanding of AI’s role in ethical decision-making and offers practical implications for designing AI-driven support systems.