To evaluate the perceived usefulness of a flipped-classroom model that integrated podcasts generated using an artificial intelligence (AI) tool in undergraduate dental education. This within-subject, quasi-experimental study was conducted at a Swiss dental school. A cohort of 49 clinical-year dental students participated in eight conventional seminars and two flippedclassroom seminars, each covering a distinct clinical topic. In both formats, a subset of 5-6 students acted as instructors under the supervision of residents or faculty members. The flipped-classroom model required students to engage with AI-generated podcast summaries of research articles before in-class discussions. These podcasts were created using NotebookLM, an online platform powered by a large language model. Surveys using Likert-type scales assessed student and student instructor satisfaction, engagement, and learning preferences. The surveys obtained response rates of ≥93.9%. The flippedclassroom model achieved high satisfaction levels among students. Compared with conventional seminars, students reported increased engagement and comprehension in the flipped-classroom format. Student instructors valued the ease of preparation and the enhanced interactivity of in-class sessions. AI-generated podcasts were rated as effective preclass resources. The integration of a flipped-classroom model with AI-generated podcast resources seemed to improve student self-perception of engagement, comprehension, and satisfaction in dental education. This instructional approach may also foster critical appraisal skills for AIgenerated content-an increasingly essential competence in academic and professional practice as AI tools gain prominence.
Eggmann et al. (Wed,) studied this question.