Rethinking Writing Assessment in the Age of Generative AI: Toward a Conceptual Framework for First-Year Writing

Abstract:

Emerging technologies have changed the dimensions of the first-year writing classroom. Before AI, assessment focused on traditional rubrics that emphasized arguments, rhetorical strategies, and language conventions. While this traditional checklist remains important for assessing student writing, it is also necessary to argue that the traditional rubric is insufficient for evaluating student performance in the current age of AI. We acknowledge that a rubric is not necessarily designed to detect AI use. However, there is a growing need for AI-informed assessment. A final-product rubric alone may provide insufficient evidence about the student’s writing process, decision-making, and authorship when AI is involved. An AI-informed assessment offers additional dimensions for evaluation, including student agency, AI literacy, ethical use, reflection and metacognition, and revision, to assess student writing skills and knowledge. This article argues that an AI-informed assessment is not a replacement for the existing rubric conventions but expands the assessment horizon in the AI-mediated first-year writing classroom. It identifies interrelated assessment dimensions—agency, AI literacy, ethical use, reflection, metacognition, and revision—and describes observable forms of evidence through which each dimension could be assessed. For example, agency may be examined through students’ explanations of their decisions to accept, modify, or reject AI-generated suggestions; AI literacy through their ability to evaluate AI outputs for accuracy, relevance, and rhetorical appropriateness; ethical use through transparent disclosure and responsible integration of AI; and reflection and metacognition through students’ explanations of their writing decisions and learning processes. Revision focuses on how students critically transform both their own writing and AI-generated material rather than simply accepting machine-produced text. The article argues that the use of AI itself is not the primary thing being graded. Instead, the framework assesses the student’s rhetorical and intellectual decisions surrounding AI use.