Malpractice or System Failure? Reconstructing Medical Liability for AI-Assisted Diagnostic Errors in Indonesia

Abstract

AI-assisted diagnostic technologies challenge conventional concepts of medical malpractice because patient harm may result from interactions between human judgment and algorithmic systems. When a physician relies on an incorrect AI-generated recommendation, determining whether the resulting harm constitutes medical negligence, technological failure, or both becomes legally complex. This article investigates the appropriate legal characterization and allocation of liability for AI-assisted diagnostic errors in Indonesia. Through normative legal analysis, the study examines the standards of professional medical care, negligence, institutional responsibility, product-related liability, and emerging principles of AI accountability. The research finds that applying conventional malpractice doctrines exclusively to physicians may inadequately address errors caused by defective algorithms, biased training datasets, inadequate validation, or insufficient warnings from technology providers. Conversely, imposing strict liability on developers for every clinical error may undermine innovation and fail to recognize the physician's independent professional judgment. The article proposes a layered accountability model distinguishing clinical negligence, technological defects, inadequate system governance, and shared causation. Liability should correspond to each actor's ability to prevent, identify, and mitigate the relevant risk. This approach would provide more equitable remedies for patients while establishing clearer responsibilities for physicians, hospitals, and AI developers in Indonesia's emerging digital healthcare ecosystem.