The AI Student Test: A Learning Centered Framework for Evaluating Generative AI Use in Higher Education

The rapid adoption of generative artificial intelligence in higher education unsettles established distinctions between learning and plagiarism, particularly in assignments built around summarization and paraphrase. Traditional academic integrity frameworks emphasize intent, authorship, and originality, yet these concepts translate poorly to machine‑generated text and offer limited guidance for evaluating AI‑mediated work. This article proposes the AI Student Test, a learning‑centered framework for distinguishing AI use that supports student understanding from AI use that substitutes for it in reading‑ and writing‑intensive assessment. The Test articulates four criteria: legitimate source use under academic norms, transformative engagement with material, traceability or acknowledgment of AI assistance, and contribution to defined learning outcomes, and operationalizes them through an evaluative rubric. The rubric is illustrated through common classroom scenarios, including AI‑assisted reading summaries, essay drafting, and take‑home assessments, showing how judgments shift from a focus on whether AI was used to whether submitted work evidences learning. The article concludes with implications for assignment design, academic integrity policy, and the integration of generative AI into the moral economy of the university in ways that preserve fairness, responsibility, and educational purpose.