This thesis investigates the applications of artificial intelligence, specifically Large Language Models (LLMs), in autonomous robotic systems within the healthcare sector. The primary objective of this study is to enhance robots' capabilities in natural language understanding and autonomy, as well as to improve human-robot interaction. The integration of LLMs into robotic systems has led to significant advancements in the autonomy and functionality of robots, enabling more natural interactions with patients and medical staff. The thesis presents the architecture of LLMs and the core technologies that support them, such as deep learning and neural networks, while examining techniques for integrating these models into robotic systems. Specifically, it explores LLM applications in diagnosis and treatment, robotic surgery, mental health support, and the automation of healthcare support services. Furthermore, the study analyzes technological limitations, ethical concerns, and security challenges associated with the use of these technologies. Finally, the thesis examines the prospects of LLMs in robotics, highlighting potential improvements and new applications expected to shape the healthcare sector.