Threshold concepts for writing with AI: Experimentation, expertise, agency

How does writing with generative AI transform our experiences as thinkers, learners, and communicators? This study investigates the question through threshold concepts, transformative ideas that prove troublesome yet shape disciplinary participation, previously identified in writing studies through retrospective analysis. We propose that technological disruption creates conditions for observing new concepts as they emerge within communities of practice. To test this approach, we designed an experimental “AI and Writing” course. Across two iterations (Fall 2023 and Fall 2024), 38 undergraduate students from 22 majors engaged in scaffolded challenges and self-directed projects, developing strategic approaches to AI-assisted composing. Analysis of students’ final reflections, using Meyer and Land’s characteristics as sensitizing criteria, revealed three concepts essential for productive AI engagement: writing with AI is an experimental process; writing with AI requires expertise and dialogue; and writing with AI should augment rhetorical agency. These transformations prove troublesome because they contradict expectations shaped by prior experience: better tools should require less effort, linguistic fluency signals substantive knowledge, and assistance either wholly helps or harms learning and development. The findings show how AI transforms perennial questions in writing studies and composition, and they offer insights for navigating the threshold between productive collaboration and problematic dependency.