Autonomy is a central concept in both ethics and bioethics, yet the diversity of its theoretical interpretations and practical applications has not been fully integrated into the ethics of artificial intelligence, particularly in healthcare. Despite the emphasis on autonomy in regulatory frameworks and ethical guidelines, the prevailing discourse often lacks conceptual clarity and overlooks distinctions between individualistic and relational accounts of autonomy. This chapter argues that these conceptual differences have significant implications for understanding the ethical challenges posed by AI systems used in clinical decision-making. Drawing from philosophical ethics and bioethics, this chapter proposes a methodological framework based on two axes: the contrast between individualist and relational conceptions of autonomy, and the distinction between informed consent and shared decision-making as models for operationalizing autonomy in medical practice. By applying these lenses to current debates in AI ethics, this chapter seeks to reveal how different normative assumptions lead to different interpretations of ethical risks, responsibilities, and patient agency. Furthermore, it contends that a relational understanding of autonomy can enrich the ethical assessment of AI in healthcare by foregrounding the roles of trust, responsibility, and care.