The Influence of Artificial Intelligence Autonomy on Physicians Work Outcomes in Healthcare: A Lab-in-the-Field Experiment

Abstract

Artificial intelligence (AI) has emerged as a powerful tool for complex decision-making. However, little is known about how AI-based systems should be integrated into clinical practice to improve patient and provider outcomes. We argue that AI-based systems are no longer passive tools and can assume certain responsibilities for tasks. A novel aspect of our study is the exploration of distribution of autonomy between AI and physicians, which can significantly impact medical care. Therefore, we shed light on how AI autonomy influence physician-AI interactions and work outcomes. We perform a lab-in-the-field experiment on diagnosis and reporting by physicians using explainable AI (XAI) in their workflow. We find that an autonomous XAI has positive effects on objective performance and intention-to-use in diagnosis and reporting. Further, we contribute to information systems (IS) research and practice by highlighting that reduced autonomy of physicians could lead to better work outcomes and improve medical care.