Artificial intelligence: ChatGPT as a disruptive didactic strategy in dental education.

Abstract

To evaluate the influence of ChatGPT on academic tasks performed by undergraduate dental students. Fifty-five participants completed scientific writing assignments. First, ChatGPT was utilized; subsequently, a conventional method involving the search of scientific articles was employed. Each task was preceded by a 30-min training session. The assignments were reviewed by professors, and an anonymous questionnaire was administered to the students regarding the usefulness of ChatGPT. Data were analyzed by Mann-Whitney U-test. Final scores and scores for the criteria of utilization of evidence, evaluation of arguments, and generation of alternatives achieved higher values through the traditional method than with ChatGPT (p=0.019, 0.042, 0.017, and <0.001, respectively). No differences were found between the two methods for the remaining criteria (p>0.05). A total of 64.29% of the students found ChatGPT useful, 33.33% found it very useful, and 3.38% not very useful. Regarding its application in further academic activities, 54.76% considered it useful, 40.48% found it very useful, and 4.76% not very useful. A total of 61.90% of the participants indicated that ChatGPT contributed to over 25% of their productivity, while 11.9% perceived it contributed to less than 15%. Concerning the relevance of having known ChatGPT for academic tasks, 50% found it opportune, 45.24% found it very opportune, 2.38% were unsure, and the same percentage thought it is inopportune. All students provided positive feedback. Dental students highly valued the experience of using ChatGPT for academic tasks. Nonetheless, the traditional method of searching for scientific articles yield higher scores.