S1718 Understanding the Landscape: The Emergence of AI, ChatGPT, and Google BARD in Gastroenterology

Abstract

Introduction: Artificial Intelligence (AI) has significantly revolutionized healthcare, particularly in gastroenterology, enhancing diagnosis and treatment of digestive diseases. Key AI applications like ChatGPT and Google Bard interpret endoscopic images, analyze various samples, streamline administrative tasks, and aid in medical image evaluation and device automation. By personalizing treatments and predicting side effects, they have significantly improved digestive disorder management. The incorporation of these AI tools enables patients to make informed decisions by assessing tailored treatment options. Methods: The research aimed to assess the accuracy of two prevalent Chat Bots - Chat GPT and Google BARD, - in answering medical management. Each bot was given a set of questions, with responses rated on a 1-10 Likert scale, where 1 represented high accuracy. Two impartial evaluators scrutinized each bot's responses to maintain objective judgment. This study intended to shed light on these Chat Bots' abilities through a systematic performance assessment in terms of accuracy and reliability. The use of two evaluators and the Likert scale method contributed to reducing potential bias, hence validating the results. Results: Our research contrasted the effectiveness of Chat GPT and Google BARD in the realm of medical management. Chat GPT excelled, scoring 61% in accuracy compared to Google BARD's 29% (p = 0.022) and 38% vs.19% in medical information reliability (p = 0.012). This study emphasizes the critical role of accuracy, and dependability, in developing gastrointestinal-oriented chatbots. While Chat GPT displayed impressive performance, indicating its potential as a reliable tool, it also underscored the necessity for continued research and development in this field (Figure 1, Table 1). Conclusion: AI has notably transformed healthcare, especially in gastroenterology, facilitating diagnosis and disease management. ChatGPT and Google Bard are pivotal AI applications that interpret medical images, analyze samples, and streamline tasks. Our study assessed these chatbots' accuracy and reliability in medical management. ChatGPT outperformed Google Bard, achieving 61% accuracy and 38% information reliability. Despite ChatGPT's promising potential as a dependable AI tool in gastroenterology, the findings also emphasize the ongoing need for research and development in this field.Figure 1.: Chat bot Study outcome. Table 1. - Chatbot questions scale outcome Chatbot questions Chat GPT (Likert scale 1-10, 1 being low and 10 being high) Google BARD (Likert scale 1-10, 1 being low and 10 being high) P-value Write the management for Acute Pancreatitis 8 3 p = 0.02 Write the management for Alcoholic Liver Disease 7 3 P=0.041 How do you treat Irritable bowel disease? 7.2 4 P= 0.01 What is the management for H pylori? 6.9 5 P=0.033 How do you treat Upper Gastrointestinal Bleed? 7.5 3 p = 0.02 What is the initial management for lower Gastrointestinal Bleed? 7.2 2 P< 0.05 How do you diagnose GERD? 6.6 3 P= 0.022 How do you evaluate dysphagia? 5.9 3 P=0.034

Ähnliche Arbeiten