Generative AI-Assisted Curriculum Design for Life Science Education: An Analysis of Conceptual Clarity, Cognitive Engagement, and Learning Achievement

This study examines the use of teacher-supervised generative artificial intelligence (GenAI) to design curriculum byemploying the GenAI in a reproducible dataset. The study uses a pretest/posttest control group design to evaluate aGenAI-assisted curriculum in life science. The sample consists of 140 students (70 in each group) and an 8-weekinstructional period. The indicators of effectiveness assessed the students’ understanding and included engagementand achievement. The analyses employ both qualitative and quantitative techniques. The techniques include but arenot limited to: analysis of covariance (ANCOVA), paired and independent samples tests, effect size calculations, 95%confidence intervals, Pearson product-moment correlations, and multiple regressions. The control and experimentalgroups did not differ statistically on any of the baseline measures (all p > .27). The experimental group showed moreimprovement on all measures of engagement, clarity, and achievement (mean difference = 5.35, 4.08, and 5.96,respectively). After accounting for the students’ pre-test achievement, the effect of group membership on students’achievement was statistically significant (p < .001). GenAI in teacher-mediated models of curriculum design can beeffective in engaging students in scientific practices, for example, to create a graph or determine and justify a scientificconclusion.