Learners Do Not Trust GenAI, but They Find It Useful Anyway: Pedagogical Implications of the Enduring Allure of GenAI for Learning

Abstract

This paper explores how learners integrate generative artificial intelligence (genAI) as part of their wider education. It used a constructivist grounded theory analysis of 35 former first-year writing students' experiences with AI tools. The paper reports on the results from a secondary analysis of the semi-structured interview portion of the project. This analysis suggested that there was a discrepancy in how students used genAI to support their learning: while learners demonstrated skepticism about genAI's reliability, they continued to find genAI tools alluring for learning support. The results place a spotlight about the quality of learners’ abilities to distinguish the quality of genAI’s outputs from its accessibility as a tool to support learning. In one sense, learners were able to determine when and how genAI might be best used to support their learning. In another sense, their discussions placed less emphasis on evaluating the outputs themselves. This paper will grapple with this discrepancy to consider ways that teachers could support learners as they integrate genAI in their learning processes.