Effectiveness of an Artificial Intelligence–Assisted HIV Educational Video for Improving HIV Health Literacy Among Indonesian Adolescents

Abstract

Adolescents in low- and middle-income countries continue to face substantial barriers in accessing effective HIV education, despite global efforts in prevention. This study evaluated the effectiveness of a culturally adapted HIV educational video developed using artificial intelligence (AI)-assisted video production tools. The intervention assessed in this study was the educational video itself rather than direct interaction with a generative AI system. A quasi-experimental design with pre- and posttests and a control group was used to evaluate the intervention among 60 high school students. Participants were randomly assigned to either a group that received a 10-min AI-generated educational video or a control group that received no intervention. The video was designed to improve HIV health literacy by addressing four domains: access to information, comprehension of core concepts, evaluation of risks, and application of knowledge in real-life decisions. Quantitative data were collected in May 2024 using a validated scale and analyzed with nonparametric tests and analysis of covariance. Students who received the AI-assisted educational video demonstrated significantly higher posttest HIV health literacy scores than the control group (Mann–Whitney U = 206.5, p < .001). After adjustment for baseline scores, analysis of covariance confirmed a significant group effect ( F (1,57) = 21.70, p < .001, partial η 2 = 0.28), indicating a large adjusted effect size. The findings highlight the potential of generative AI as an innovative, scalable tool for adolescent health education, particularly in resource-limited settings. This approach may offer new opportunities for integrating digital strategies into school-based health promotion programs.