Difficulty as method: A framework for humanities pedagogy in the age of generative AI

Generative AI poses a distinctive challenge for humanities pedagogy, one that existing AI-in-education frameworks do not fully address. Humanities teaching is organized around interpretive openness, productive ambiguity, close reading, and a conception of evidence as something argued for rather than merely cited. In common classroom use, generative AI favors fluency, plausibility, and resolution in ways that work against those commitments. This article proposes a conceptual framework that treats AI as material for interpretive inquiry rather than as a tool for efficiency. Drawing on rhetorical analysis, dialogic pedagogy, and theories of revision, it develops four instructional roles for AI in the humanities classroom: object of interpretation, rhetorical artifact, revisable draft, and dialogue partner. The article argues that the most productive response to AI in the humanities begins with disciplinary redesign rather than with adoption or prohibition--and that the difficulty the humanities cultivate is not an obstacle to learning but its method.