Feasibility and Challenges of Interactive AI for Traditional Chinese Medicine: An Example of ChatGPT

Abstract

ChatGPT, developed by OpenAI, is currently the largest language model with robust interactive capabilities. As a complementary alternative medicine (CAM), Traditional Chinese Medicine (TCM) represents an established medical system with a rich history and abundant clinical experience. TCM is an empirical medicine, the process of which is analogous to ChatGPT's learning and development model. In TCM, inquiry is a relatively objective way of traditional syndrome differentiation. Although various artificial intelligence systems related to TCM consultation exist, their interactive abilities remain limited. As such, we standardized the primary complaint and instructed ChatGPT to simulate a TCM practitioner, conducting three comprehensive inquiry tests. The results yielded unexpected conclusions, revealing that ChatGPT could simulate a TCM practitioner's inquiry with patients, confirming its potential in the field of TCM inquiry. However, current applications still pose certain limitations and risks. Hence, to integrate ChatGPT-like language models with traditional TCM AI to establish an associative mode that can facilitate TCM diagnosis and treatment with more convenience and standardization is crucial, yet at the same time, it should be treated very carefully.