Author(s)

Fatmah Saif Obaid Alhefeiti¹, Mostafa Ezzat², Nesrine Ali Abd El Azim¹, Hesham A. Hefty²

¹Department of Information Technology, FGSSR, Cairo University, Cairo, Egypt.
²Department of Computer Science, FGSSR, Cairo University, Cairo, Egypt.

Artificial intelligence (AI) is progressively influencing various fields, with its impact on healthcare being particularly significant. The Transformer neural network architecture, initially developed for a range of Natural Language Processing (NLP) tasks, is now being adapted for multiple applications in the healthcare sector. This study employs a systematic literature review (SLR) to evaluate research published between 2017 and 2024, focusing on five key research questions to interpret and analyze the relevant literature. This comparative analysis examines the advancements and effectiveness of AI-driven chatbots in healthcare, specifically highlighting the use of the Transformer architecture in analyzing diverse healthcare data types, including clinical NLP, medical imaging, and interactions on social media. We also discuss studies that have leveraged Transformer models to generate surgical instructions and predict adverse outcomes in critical care environments post-surgery. Furthermore, we propose a framework for future advancements that incorporates user feedback, ethical considerations, and technological innovations to develop more robust and reliable AI healthcare solutions. This comparative study contributes a framework for future developments that incorporates user feedback, ethical considerations, and technological innovations, aiming to enhance the reliability of AI healthcare solutions. Ultimately, our findings highlight the transformative potential of AI chatbots in healthcare and emphasize the need for ongoing research to address existing challenges.

Medical Chatbot, NLP, Artificial Intelligence (AI)

Alhefeiti, F. , Ezzat, M. , Azim, N. and Hefty, H. (2025) A Comparative Study of AI-Powered Chatbot for Health Care. Journal of Computer and Communications, 13, 48-66. doi: 10.4236/jcc.2025.137003.