The development of Artificial Intelligence (AI) technology in healthcare brings two legal consequences: increased efficiency and accuracy of diagnosis, on the one hand, and the potential for AI-based medical crimes, such as misdiagnosis, negligence in operating systems, and misuse of patient health data. The lack of regulations specifically addressing legal liability for errors involving AI creates legal uncertainty, both for patients as victims and for medical personnel who may be criminalized for system errors beyond their control. This study aims to examine the characteristics of AI-based medical crimes and analyze the relevance of a restorative justice approach as a resolution mechanism that can achieve balanced legal protection for patients and medical personnel. This study uses a normative juridical method with a statutory approach and a conceptual approach. The research findings indicate that restorative justice, as mandated by Article 306 and Article 310 of Law Number 17 of 2023 concerning Health, can be applied to AI-based medical crimes involving minor negligence (culpa levis), but is inappropriate for cases involving intentional negligence (dolus) or gross negligence (culpa lata). The application of restorative justice requires strengthening of technical regulations regarding audit standards, certification, and the division of responsibilities between medical personnel, healthcare facilities, and AI system developers to ensure that the goals of legal certainty, justice, and benefit can be achieved proportionally.