Ideating Worlds with ChatGPT: Critical Inquiries into the Impact of Artificial Intelligence on Narrative Composing

• In Event: Speculative Futures with and Against AI: Youth Worldbuilding as Sociotechnical Intervention

Abstract

As educators and scholars make sense of artificial intelligence (AI) and its role within learning communities, there is a breadth of research that investigates the instructional benefits and challenges of students’ AI usage (Chan et al., 2023; Duong et al., 2023; Wu et al., 2024). The latter studies work to determine the pedagogical utility of AI platforms as they pertain to learning across academic disciplines. A complementary line of research inquires into the biases of AI tools and their sociopolitical and environmental impact around the globe (Benjamin, 2019; Crawford, 2021). Threading the latter two programs of research together, Vauhini Vara (2025) argues that conversations praising the utility of AI platforms can sometimes work to insidiously perpetuate their well-documented escalation of environmental destruction (Li et al., 2023) and oppressive labor practices (Hao, 2025). Rather than work toward a simplistic vision of AI utility in schools, the present study sought to engage a group of pre-service teachers (PSTs) in a series of critical encounters with ChatGPT to understand how it advanced and constrained their narrative composing practices. I asked: 1) How did a group of 28 pre-service teachers in a literacy master’s program use inquiry-based prompts to solicit narrative ideation feedback from ChatGPT?, 2) How did the PSTs’ interactions with ChatGPT impact their conceptions of AI as a tool for narrative feedback? Thinking across dimensions of compositional practice and human/machine relationality, this paper seeks to nuance the affordances and limits of AI usage in ways that underscore the specific genius of human-generated narrative thinking.
Data for this study was generated in a graduate Teaching YA Literature course during the Spring 2025 semester. The final project of the course – titled, You Be The Author – invited PST participants to develop a short YA narrative across multiple iterations of feedback from ChatGPT. Throughout the process, grounded in constructivist approaches (Vygotsky, 1978) to literacy learning, the PSTs critically reflected on how their interactions with ChatGPT impacted their narrative writing. In sum, the PSTs generated 28 short story narratives, across three iterations of revisions, based on their interactions with ChatGPT. In addition to the 28 narratives, Author 1 conducted three retrospective design interviews (Dalton et al., 2015) with each PST to understand their writing process and relationships with ChatGPT.
Findings for this study highlight a divergence in PSTs’ opinions about ChatGPT as a source of narrative feedback. While some PSTs argued that ChatGPT’s open-ended questions about their narrative helped them elaborate their stories, others noted a variety of critical limitations to ChatGPT’s feedback. For example, one PST noted that ChatGPT refrained from asking questions about protagonists’ gendered and racialized identities in ways that could have informed the development of her narrative. Another PST noted how ChatGPT’s feedback didn’t draw upon nuanced understandings and explorations of his familial relationships in the ways that a close friend might. Taken together, the critiques of ChatGPT’s utility – particularly those centered around inclusive, justice-oriented, and relational approaches to narrative writing – highlighted aspects of the writing process that ground storytelling in the human experience.

Author

• Alex Corbitt, Syracuse University