Exploring Teachers’ Integration of Generative Artificial Intelligence in Assessment Design

Abstract

This study explores the impact of integrating generative artificial intelligence into the design of alternative assessments on teachers’ epistemic fluency and cognitive skills. Using a mixed-methods approach,14 Taiwanese teachers participated in a pre- and post-test intervention supported by qualitative inquiry. Quantitative results from paired t-tests showed significant improvements in both critical and creative thinking. Epistemic fluency showed a positive upward trend but did not reach statistical significance. Qualitative content analysis further elucidated these changes, revealing that teachers exhibited high epistemic fluency through sophisticated, multi-layered verification frameworks to counter AI-generated information. The findings suggest that while AI acts as a catalyst for idea inspiration and a rigorous tool for honing judgment, its efficacy is primarily determined by the teacher’s professional agency.