Effectiveness of an AI-Based Educational Video Generator in Enhancing Oral Health Knowledge and Behavior Among Children: A Randomized Controlled Trial

Background: Artificial intelligence (AI) is a new innovation that presents an opportunity for oral health education in children. Objective: To evaluate the effectiveness of an AI-assisted educational video versus standardized in-person oral health education for children regarding oral health knowledge and behaviors. Methods: This randomized controlled trial included 300 children aged 10-12 years (n=150 per group), both boys and girls, who were divided into AI video and standardized face-to-face oral health education groups. Baseline and 2-week follow-up evaluations were conducted for knowledge scores, four behavioral items, and the Silness and Löe Plaque Index. The internal consistency of the knowledge scale was checked (Cronbach's α=0.79). Wilcoxon signed-rank and Mann–Whitney U tests were the primary non-parametric tests used for statistical evaluations, with a median, (IQR), and effect size r provided. Chi-square tests with a Bonferroni adjustment at α=0.0125 were used to evaluate behavioral items. Results: There were significant enhancements in intra-group knowledge (p<0.0001 for both groups; r=0.663 for AI; r=0.345 for Control). The clinical efficacy was similar for both strategies with significant reduction of intra-group plaque (AI: p< 0.001, r= 0.425; Control: p< 0.001, r= 0.331), but no inter-group difference at the end of the intervention (p= 0.608; r= 0.030). In the conventional group, the utilization of dental floss was significantly improved (p=0.0056), and no other behaviors showed significant improvement across groups. Conclusions: The AI-assisted educational video demonstrated better knowledge retention than the standardized face-to-face oral health education, achieved similar clinical plaque reduction, and can be easily scaled for pediatric preventive dentistry.