Beyond the Black Box: Harmonizing AI-Driven Medicine with Ethical-Legal Imperatives to Safeguard Human Autonomy and Accountability

The integration of artificial intelligence (AI) is profoundly reshaping healthcare, enhancing efficiency while introducing complex challenges in algorithmic transparency, legal accountability, and the safeguarding of fundamental rights. The convergence of AI and medicine creates significant ethical-legal dilemmas concerning liability for algorithmic errors, patient self-determination, and the evolving role of healthcare professionals. This paper addresses these issues by proposing a novel, functional liability framework based on the AI system's Level of Autonomy (LoA). Using an interdisciplinary methodology, it analyzes the limits of existing legal paradigms and the EU's regulatory ecosystem. The proposed LoA framework offers a structured approach to distribute responsibility among developers, healthcare institutions, and clinicians, harmonizing technological advancement with human-centric values. The paper illustrates the framework's practical application through case studies and outlines a path for implementation, advocating for solutions like risk-based insurance and no-fault compensation schemes for higher autonomy levels.