Use and Perceptions of Large Language Models Among Dental Students: Implications for Dental Education

Abstract

PURPOSE: This study aimed to quantify the prevalence and use cases of large language models (LLMs) among dental students, estimate perceived usefulness (learning enhancement; time saved), characterize concerns (accuracy, ethics/integrity, appropriateness), and derive actionable implications for curriculum and clinical training. MATERIALS AND METHODS: A cross-sectional survey was administered to dental students and postgraduate trainees at a single institution across all years. The questionnaire included items on familiarity with LLMs, frequency and purpose of use, perceived helpfulness, time-saving benefits, concerns, and preferences for integration into dental education. Descriptive statistics, phase comparisons, and logistic regression models were used to analyze quantitative data, while free-text responses were examined thematically. RESULTS: Out of the 156 respondents, 80.7% reported using LLMs, most often for concept clarification (59%) and exam preparation (36%). Nearly half (49%) found LLMs helpful, and 62% reported time-saving benefits. Concerns were expressed by 23%, mainly regarding accuracy, discipline-specific limitations, and ethical implications. Phase differences were small and largely not significant, suggesting that perceived usefulness and reliability outweighed the level of training. Logistic regression showed that students rating LLMs as not helpful or neutral had 5.6 times higher odds of expressing curricular concerns. In contrast, those reporting time saved were more likely to use LLMs frequently. Thematic analysis highlighted tensions between efficiency and accuracy, with students valuing time-saving but worrying about inaccurate information. CONCLUSION: Our findings indicate that LLMs are already embedded in dental students' study practices, with the greatest value in non-clinical learning tasks. Structured integration with verification and the need for faculty oversight are essential to maximize benefits while addressing concerns about accuracy and professionalism.

Ähnliche Arbeiten