The effect of a human-GAI dialogue mode on improving college students’ questioning competence in digital reading

Abstract Students’ questioning competence, which is strong related to higher-order cognitive processes such as critical thinking, creativity, and problem-solving, is crucial to develop during reading. Drawing on dialogue learning theory, this study examined the effects of a human-GenAI dialogue learning mode on college students’ post-reading questioning in digital reading. An instructional quasi-experiment with an in-subject design was conducted among 112 college students, who completed reading tasks under both a general reading mode and a human-GenAI dialogue mode supported by ERNIE Bot. Students’ post-reading questions were analyzed in terms of quantity, cognitive level, external structure, and overall quality. Students’ output in the human-GenAI dialogue were analyzed through lag sequential analysis (LSA), complemented by representative dialogue excerpts. It was found that the application of the human-GenAI dialogue learning mode significantly improved students’ questioning behaviors and competence during reading. The lag sequence analysis of students’ strategies use in human-GenAI dialogue reveals that different strategies used in human-GenAI dialogue can affect students’ questioning performance. Students who have a higher level of self-regulation, are good at reflecting and adjusting their dialogue strategies tend to raise higher-quality questions. Overall, the findings suggest that the human-GenAI dialogue mode can serve as a scaffolded environment that supports higher-quality post-reading questioning. It may help understand how human-GenAI dialogue may support reading-related inquiry and offers implications for designing AI-supported reading activities that foster more purposeful and reflective questioning.