When AI Writes and AI Grades

Abstract

Generative artificial intelligence increasingly mediates both student writing and educational assessment, creating recursive workflows in which machine-generated discourse may be evaluated by other machine systems. This study examines the pedagogical, ethical, and political-economic implications of that development through qualitative discourse analysis and heuristic comparison. Two purposively selected, high-engagement Moltbook exchanges are analyzed alongside a researcher-constructed simulation in which ChatGPT generates a first-year writing essay and rubric-based feedback on that essay. Findings suggest that AI-mediated writing and assessment can preserve recognizable forms of authorship, dialogue, and evaluation while redistributing authorship, judgment, and responsibility across users, models, rubrics, platforms, and institutions. The analysis distinguishes criteria-based feedback from situated pedagogical judgment and treats technofeudalism as an interpretive lens rather than an empirical conclusion. The article argues for accountable human review, process-oriented assessment, transparent data governance, careful procurement, and meaningful alternatives when educational work is routed through third-party AI systems in consequential institutional settings.