Exploring Jordanian medical students' perceptions and concerns about ChatGPT in medical education: a cross-sectional study.

Abstract

Background: The integration of Artificial Intelligence (AI) tools like ChatGPT into medical education is expanding, offering benefits such as efficient information synthesis. However, concerns about the accuracy, reliability, and proper use of these tools persist. Understanding medical students' perceptions of ChatGPT is crucial for optimising its use in educational settings.

Objectives: To evaluate how medical students perceive ChatGPT for educational purposes and to assess its perceived advantages and disadvantages.

Methods: A cross-sectional study was carried out using a questionnaire with five main domains to explore Jordanian medical students' perceptions, practices, and concerns regarding the ChatGPT. This study was conducted from May to July, 2023, and the data were collected using the convenience sampling technique through Google Forms shared within medical students' Facebook groups. Descriptive statistics summarised participant demographics, while logistic regression identified factors influencing ChatGPT usage. Variables with a P-value ≤ 0.05 in multiple regression were considered statistically significant.

Results: Nearly two-thirds (N = 136, 61.5%) claimed to have knowledge of AI but not in clinical settings. Most participants (88.5%, N = 216) were aware of ChatGPT, with 86.9% (N = 212) agreeing that 'Medical students can benefit from using ChatGPT.' Additionally, 83.2% (N = 203) felt that 'ChatGPT helps students quickly and easily summarize complex information.' Conversely, 78.3% (N = 191) expressed concerns about ChatGPT's potential inaccuracies, with accuracy and reliability cited as primary concerns. Multiple logistic regression showed that younger students (OR = 0.902, P = 0.025) and those with lower proficiency (OR = 0.487, P = 0.007) used ChatGPT more frequently than others.

Conclusion: Although the use of the ChatGPT could be more beneficial for aiding students in developing medical knowledge, evidence-based academic regulations should guide its use. Future research should be conducted to examine the enablers and barriers to ChatGPT use in medical education.