Distinguishing AI-generated and real images: implications for media literacy instruction

Abstract

This study aimed to assess the detection accuracy and usefulness of determinants for distinguishing artificial intelligence (AI)-generated and real images among senior high school students in Western Visayas during the academic year 2025–2026. It examined the significant differences in the two mentioned constructs across students’ demographic profiles. It employed a quantitative design- the descriptive-comparative approach. A random sample of 104 students was involved, using a validated, reliability-tested 33-item researcher-made questionnaire. Descriptive and comparative statistical analyses were employed. Mean, standard deviation (SD), Mann-Whitney U test, and Kruskal-Wallis H test were used. The findings revealed that the students demonstrated a moderate level of detection accuracy (M=5.16, SD=1.61). The students rated the usefulness of the determinants as helpful (M=3.78, SD=0.74). Additionally, it was found that there is a significant difference in detection accuracy when grouped by school type [U=965.500, p=0.011] compared to its counterparts. Significant differences were found in the usefulness of determinants for distinguishing AI from real images when grouped by type of school [U=815.000, p=0.001] and frequency of use of AI [H(2)=11.989, p=0.002]. The findings of the study may provide teachers with insights into developing instructional strategies and interventions to enhance students’ critical thinking and digital literacy skills.