When AI Use Is Not Simply Generation: The GOPA Framework as a Shared Vocabulary for Human-Centered AI in Education

Abstract

Generative artificial intelligence (AI) is rapidly reshaping educational practice by enabling the production of instructional materials, assessments, feedback, explanations, and learning supports with unprecedented speed. However, discussions of AI in education often rely on broad labels such as “AI-generated” or “AI-assisted,” which can obscure important differences in how cognitive and epistemic responsibilities are distributed between humans and AI systems. This paper introduces the GOPA Framework; Generation, Organization, Personalization, and Analysis as a shared vocabulary for distinguishing AI-Centered Production from Human-Centered Augmentation in educational contexts. Grounded in Human-in-the-Loop perspectives, the Curated Authorship Model, cognitive offloading research, epistemic cognition, and Human–AI augmentation, the framework examines where responsibility for interpretation, adaptation, evaluation, and judgment resides within AI-supported work. Generation describes uses in which AI assumes primary responsibility for producing content and meaning, while Organization, Personalization, and Analysis describe forms of AI use that support human thinking while preserving human responsibility for meaning making. The paper argues that educationally meaningful AI use should not be evaluated only by efficiency or output quality, but by whether learners and educators remain active drivers in constructing, adapting, evaluating, and interpreting knowledge. By prioritizing agency, authorship, accountability, epistemic authority, and productive cognitive friction, the GOPA Framework offers educators a practical language for examining AI use in teaching, learning, writing, and knowledge construction.