Creative Thinking in Human-AI Dialogue: A Qualitative Study of History Learning with Generative AI

Abstract

The rapid development of generative artificial intelligence (generative AI) has introduced new possibilities for transforming how students interact with information, construct knowledge, and engage in inquiry-based learning. However, the educational value of generative AI depends not only on its capacity to generate content but also on how learners critically engage with and reflect upon AI-generated responses. In history education, where learning involves questioning evidence, interpreting the past, and constructing reasoned explanations, understanding how students think through interactions with generative AI has become an important research concern. Existing research has largely focused on the instructional applications and learning outcomes of generative AI, with limited empirical attention to how creative thinking is enacted through sustained human-AI dialogue during historical inquiry. This qualitative study explored how middle school students engaged with the generative AI platform Doubao during historical inquiry and how creative thinking was enacted throughout these interactions. Participants were ten eighth-grade students from a middle school in Beijing, China. Data were collected from multiple sources, including student–AI dialogue records, classroom observations and researcher field notes, and semi-structured interviews, and analysed using Braun and Clarke's thematic analysis. The analysis revealed that students' interactions with generative AI evolved from information seeking towards dialogic exploration and critical negotiation. Across these interactions, creative thinking was progressively enacted as students refined historical questions, evaluated historical evidence, considered multiple perspectives, revised historical explanations, and constructed personal historical interpretations. Rather than being used primarily for information retrieval, generative AI served as a dialogic partner through which students externalised, examined, and progressively refined their historical understanding. By conceptualising creative thinking as an observable process enacted through human-AI dialogue rather than merely a cognitive outcome, this study extends current understandings of how creative thinking can be examined in generative AI-supported history learning.

Keywords: generative AI; creative thinking; human-AI dialogue; history education; middle school student

DOI: 10.7176/JEP/17-7-05

Publication date: July 30^th 2026