Designing with Notebooklm: An Autoethnographic Discourse Analysis of Human-AI Curriculum Design

Abstract

Objective: This study examines how the linguistic choices made by a human curriculum designer and Google’s NotebookLM constructed their roles and relationship during their collaborative curriculum design, and what these patterns revealed about the distribution and mobilization of human and technological resources. Methods: This work adopts an autoethnographic design and analyzes a two-month written interaction between an experienced curriculum designer and NotebookLM during the redesign of an undergraduate education class in the UK. The corpus comprised 367 interactional turns and 50,832 words. The interaction was analyzed discursively through Systemic Functional Linguistics and Appraisal Theory. Brown’s (2011) Design Capacity for Enactment framework was subsequently employed to interpret the 48 curriculum-design episodes identified in the corpus. Results: The interaction was collaborative and increasingly personalized, but asymmetrical in design authority. NotebookLM contributed extensive textual production, elaboration, source synthesis, and strongly affirmative language, while the Human Designer expressed pedagogical intentions, evaluated proposals, controlled sequencing, and determined what should be accepted, modified, or rejected. The design episodes showed that adaptation was the most common form of resource appropriation, while instances of rejection provided additional evidence of human design authority. Discussion: The findings suggest that Generative AI can become a substantive curriculum-design resource without displacing educators’ role as designers. NotebookLM expanded the range of candidate curricular representations, but their relevance depended on human evaluation, contextualization, and transformation. Linguistic alignment and extensive AI production did not necessarily correspond with design authority or successful implementation, reinforcing the need for continued professional judgment in Generative AI-supported curriculum design.