https://orcid.org/0000-0002-4095-5253
Extract
Recent developments in large language models (LLMs), especially the representative product ChatGPT, have created new opportunities in the medical field. We have noticed numerous studies by medical researchers envisioning ChatGPT performing complex functions like clinical diagnosis and writing medical papers. However, these applications are currently impractical or cannot be used responsibly and fully trusted. We have observed that your recent publications concerning about patient consultation [1, 2] and medical education [3], which we believe are the most effective and trustworthy applications in medicine for current LLMs. We are particularly optimistic about the role of LLMs in enhancing medical communication. As social science researchers studying medical AI, we would like to discuss this issue mainly from the following two perspectives.
On one hand, a major advantage of LLMs is their ability to perform cross-cultural language translation [4]. Many users have praised ChatGPT for its “authentic” translations. By “authentic,” we mean translations that not only use words and grammar similar to native speakers but also consider cultural background, customs, and subtle nuances in context [5]. LLMs can achieve dynamic equivalence and transcreation by adjusting for cultural differences, such as modifying technical terminology and providing locally relevant examples. This makes translations not only accurate but also natural and relevant to the target audience, thereby promoting cross-cultural communication and understanding. For example, in Traditional Chinese Medicine, concepts like “Qi,”(气) “Meridians,”(经络) and “Yin and Yang”() are foundational. Chinese people also use these concepts in daily life to describe their symptoms, but they lack direct Western equivalents. Another term, “shang huo,” (上火) describes symptoms like sore throat and mouth ulcers. Even Western-trained doctors in China frequently use it, but there is no precise Western counterpart. LLMs can accurately translate and explain these concepts, enhancing communication between patients and doctors from different cultural backgrounds, as well as between patients and chatbots, and even among medical professionals.
