Cognitive Externalization and Generative AI in Higher Education

Abstract

The spread of generative artificial intelligence (GenAI) in higher education calls into question the adequacy of earlier conceptions of the digitalization of learning, since it affects not only educational infrastructure but also the cognitive operations underlying learning activity itself. The aim of the article is to interpret GenAI as a new wave of cognitive externalization and to determine the implications of this interpretation for learning design, process-based assessment, and learner agency in higher education. Cognitive externalization is defined as the transfer of some operations of memory, fixation, semantic processing, structuring, and preparation of decisions into an external symbolic or socio-technical environment while preserving the need for human control, interpretation, and responsibility or, in some cases, without actual human control when delegated operations are not consciously verified. The study employs an integrative problem-oriented analytical literature review based on a comparison of scholarship on writing as a cognitive technology and research on the impact of GenAI on university teaching and learning. The scientific novelty of the article lies in interpreting GenAI as a new wave of cognitive externalization and in theoretically grounding the previously proposed PPAIR model as a didactic response to the transformation of learning activity. It is shown that the status of a “wave” is determined not by the mere emergence of a new technology but by changes in basic cognitive operations, the object of pedagogical control, and forms of assessment. Unlike writing, which historically enabled the external fixation of memory and the accumulation of knowledge, GenAI externalizes operations of linguistic and semantic processing, the generation of alternatives, and the preparation of decisions. This changes the architecture of learning activity, reduces the diagnostic reliability of text- centric forms of assessment, and intensifies the problem of learner agency. As a process-oriented didactic response, the article theoretically grounds the PPAIR model – Problem – Prompt – AI Generation – Iterate – Reflect – which is aimed at capturing not only the final product but also the student’s trajectory of working with a generative model. It is concluded that the integration of GenAI into university education requires reconsideration of teaching tools, criteria for validating learning outcomes, and institutional foundations of academic responsibility. The practical significance of the article lies in substantiating a shift from a prohibitive logic toward the managed integration of GenAI into the educational process.