In postdigital schooling, platforms and AI reduce friction and make it increasingly rational for students to bypass demanding work. A salient case is reading long, linguistically and conceptually challenging literary classics—an educational practice whose benefits are slow, indirect, and difficult to measure. Large language models (LLMs) and summarisation tools intensify this tendency by turning texts into extractable information and rewarding fluent “hyperreal” performances of reading over sustained interpretive engagement. This conceptual paper argues that the key leadership problem is no longer only whether students comprehend what they read, but how schools can justify and sustain slow, effortful reading under LLM-rich conditions.

The paper integrates (1) experiential and interpretive theories of reading (Felski; Calvino), (2) research on deep reading in digital environments, (3) digital social reading as a motivational and interpretive resource, and (4) ethical and governance frameworks for GenAI in education (UNESCO, OECD, EU and national guidance). It develops principles for “AI-supported slow reading,” positioning AI as metacognitive scaffolding that sustains productive difficulty by supporting attention, questioning, monitoring, and interpretation without substituting for the act of reading.
The framework specifies boundaries between scaffold and substitute uses, task design criteria for productive difficulty, AI-supported routines for annotation and revision (e.g., post-reading Socratic prompts), process-based evidence structures, school-level transparency and shared rules, and equity-oriented differentiation safeguards. Targeting secondary education, the paper offers a principled basis for instructional design, curriculum work, assessment alignment, and leadership decisions that revalue slow engagement over fast outputs in postdigital literature education.