ARTIFICIAL INTELLIGENCE CHATBOTS IN PATIENT EDUCATION: OPPORTUNITIES, RISKS, AND ETHICAL CHALLENGES

Artificial intelligence chatbots and large language models are increasingly used to obtain, simplify, and personalize medical information for patients. This narrative medical review synthesizes 40 recent studies on chatbot-supported patient education across emergency medicine, cardiology, oncology, ophthalmology, gastroenterology, dermatology, orthopedics, otolaryngology, urology, pain medicine, nursing, and health-literacy research. The review evaluates opportunities, risks, and ethical challenges associated with patient-facing AI communication. The evidence suggests that chatbots can generate fluent explanations, improve access to plain-language materials, support appointment preparation, and reduce the workload of drafting educational content. However, important limitations recur across the literature, including excessive reading level, incomplete or unsafe advice, weak source transparency, fabricated references, model variability, language-related inequity, and patient overtrust in fluent but unverified answers. Ethically acceptable implementation should therefore treat chatbots as supervised communication-support tools rather than autonomous medical educators. Clinical use should require clinician-approved source material, explicit AI disclosure, readability testing, escalation instructions, model-version documentation, multilingual validation, and post-deployment monitoring. Under such safeguards, AI chatbots may strengthen patient education, but unsupervised deployment may intensify existing problems in online medical information.