Trust in Aligned AI Decision Makers

As artificial intelligence (AI) systems take on increasingly complex decision-making roles, aligning AI decision makers with human values becomes critical to fostering trust and enabling effective delegation. This study examines whether AI decision makers (ADMs) that are aligned with human attributes earn greater trust and delegation preference in high-stakes medical triage scenarios. Using a combination of attribute assessments, AI training, and human evaluations, we found that alignment significantly influences trust and delegation for most but not all attributes. These findings provide empirical evidence supporting the hypothesis that AI alignment enhances trust and delegation, offering insights into the future design of AI-assisted decision-making systems.