Comparison between ChatGPT 3.0 and Google Gemini Regarding Medicine Knowledge

Abstract

<title>Abstract</title> ChatGPT and Gemini AI are two of the most advanced and enhanced large language models widely used worldwide for various purposes. These models are built to facilitate human civilization with their generative capability to produce solutions and suggestions for different purposes, in a human-like conversation type with predictive texts. This study aimed to identify the potential differences between these two models in the case of possessing medical knowledge. A set of multiple-choice questions (MCQ) was adapted from a medicine textbook, and the correct answers were identified by matching the answers in the textbook and a medical expert. Then both of the models were asked to identify the correct answers from the options given to them. They were scored based on their ability to identify the correct answers. The findings revealed that both AI models possess significantly less knowledge of different disease domains and are not sufficiently reliable for medical assistance, though ChatGPT 3.5 possesses slightly better knowledge than its counterpart Google Gemini. The developers should focus on these models to make them more reliable in medical education so that our medical students and doctors can utilize the full potential of Artificial Intelligence in their medical lives for both learning and application.

Ähnliche Arbeiten