Late Breaking Abstract - Smart Talk for Allergies: Unveiling an AI Chatbot for Healthier Homes

<bold>Background:</bold> The prevalence of alergic diseases has been steadily increasing, highlighting the importance of effective a lergen management. However, patients often encounter misinformation about a lergen management on the internet. This underscores the need for a specialized large language model (LLM)-based chatbot, tailored for educating patients specifica ly about a lergen management. <bold>Method</bold>: We developed a chatbot based on the OpenAI assistant utilizing the GPT-4 model (gpt4-1106-preview). The Reference Augmented Generation (RAG) process included key resources such as Middletone's Alergy chapters on aerobiology, indoor a lergens, and a lergen control, along with the 3rd edition of the Textbook of Asthma and Alergic Disease. Additionaly, investigative reports on major indoor a lergens conducted by third-year medical students at Chungbuk National University were incorporated into the RAG. <bold>Results:</bold> The LLM-based chatbot for alergen management was developed using three models: OpenAI's GPT assistant, GPTs, and the no-coding tool FlowiseAI. These chatbots are accessible at designated below QR code. Figures 1-3 in our submission demonstrate the chatbot's responses to queries about environmental management for common alergens like dust mites and cats. <bold>Conclusion</bold>: Management of key indoor alergens, such as house dust mites and pet dander, is crucial for contro ling and preventing a lergic diseases. We have developed an LLM based chatbot focused on these alergens. Although the chatbot was designed to reference textbooks to reduce halucinations, further research into reducing halucinations through human reinforcement learning and fine-tuning based on these interactions is necessary.