Flipped Classroom With Artificial Intelligence: Educational Effectiveness of Combining Voice-Over Presentations and AI

Background: Most theorists and medical educators agree that a curriculum rich in active learning (AL) strategies, such as a flipped classroom, is superior to passive listening to promote better retention and application of new knowledge. Although AL multimodal teaching strategies have been considered the most effective, including online virtual teaching, voice-over pre-recorded lectures, and, more recently, the addition of artificial intelligence (AI), data on the effectiveness of these methods in medical education is scarce. The present educational research study examined the effectiveness of voice-over-style lectures and AI in facilitating learning outcomes as assessed by test scores after participating in basic science lectures in a medical school setting. Methods: Participating students were divided equally into two educational strategy groups: slide decks only traditional way (PPT) or PPT plus AI (PPT+AI) platform (edYOU; Los Angeles, CA, USA). The PPT+AI group comprised the PPT with narration and real-time interaction with an AI being personalized, which leverages natural language processing to tailor customized conversations to each student’s current knowledge. Students in the two groups were asked to participate in a formative quiz (not reflective of their academic evaluations) to answer questions relevant to voice-over lectures (PPT and PPT+AI). The statistical strategy for conducting quiz item analysis included item difficulty, item discrimination, and point-biserial correlation R. A student's T-test was conducted to compare the two strategies’ effectiveness via test scores. A priori, an alpha level of 0.05 was considered significant. Results: Data are presented as mean ± s.e.m.; Cohen’s d. A total of 42 (n=21 in each group) students participated in the study. Students using PPT+AI obtained statistically significant (P <0.043; d = .54) higher quiz scores under challenging questions and less time spent in lectures (54.1 ± 14.3 hrs.) in the PPT+AI group (P <0.001; d = 1.17) compared with the PPT group. Conclusions: The PPT+AI strategy could be the difference between a pass and a fail, as the PPT+AI strategy is particularly efficient in improving difficult question test scores. At the same time, students may learn the material in less time (efficiency). Research on the application of AI as part of educational strategies for improving satirized test scores, including boards, is warranted. The present study is part of the necessary early steps to better understand the impact of AI as an educational strategy for improving educational outcomes.