The fast development of telemedicine has shown the value of medical chatbots in facilitating interaction between patients and doctors. This study examines ChatDoctor, a health chatbot optimized with LLaMA, to determine the medical diagnosis suggested by Artificial intelligence compared to a physician’s diagnosis. We employ 100,000+ user reviews and health chatbot conversations, applying Natural Language Processing (NLP) and Large Language Models (LLMs), to investigate patient interactions. Findings indicate that although chatbots improve access, patient engagement, and convenience, they also have accuracy and trust issues. Topic modeling identifies a contradiction in user experience, where positive comments highlight chatbot support and negative comments refer to response reliability problems. This research contributes to the Information Systems literature by demonstrating how LLMs enhance telehealth and provide valuable insights for improving chatbot design and patient-centered AI solutions. The results emphasize the need for sophisticated AI models that balance efficiency, clinical accuracy, and user demands.