Authors: Ghare Vaishnavi Shahaji, Prof. C. A. Kolse, Prof. M. C. Kolse, Prof. S. K. Said
Abstract: Healthcare systems face long waiting times, shortages of clinicians and limited access to reliable health information, particularly in rural and underserved areas. AI chatbots, conversational agents that interact with patients through text or voice, can answer health questions, check symptoms, remind patients about medication and appointments, and provide basic mental health support around the clock. This paper reviews AI chatbots in healthcare. It traces their development from rule-based dialogue systems to agents powered by natural language understanding and, most recently, large language models trained or adapted on medical knowledge. Evidence from systematic reviews, randomized trials of mental-health chatbots, audits of online symptom checkers and comparisons of chatbot and physician responses is examined, together with studies of patient acceptance and of how assistants respond to crisis questions. A five-layer architecture comprising the user interface, language understanding, a grounded medical knowledge and reasoning layer, a safety and escalation layer and integration with health records is proposed. Research gaps, objectives, a comparison of chatbot types and challenges including accuracy, hallucination, privacy, liability, language diversity and digital literacy are discussed. The review concludes that healthcare chatbots are valuable as supervised, well-grounded assistants rather than replacements for clinicians.
