Regulation of Academic Integrity in the Use of Generative Neural Networks within the Student Learning Outcomes Assessment System

Abstract

The active integration of generative neural networks into the educational process is reshaping traditional models of learning outcome assessment and creating new challenges for ensuring academic integrity. The ability of digital tools to generate texts, calculations, and analytical responses complicates the determination of a student’s individual contribution. This necessitates an update of assessment principles and the development of effective mechanisms for their implementation. The aim of the article is to substantiate the principles of ensuring academic integrity when using generative neural networks in the assessment of learning outcomes, as well as to identify ways to mitigate the risks of dishonest behavior. The study employs a complex of interconnected methods. Specifically, theoretical generalization was used to systematize scientific sources, comparative analysis was applied to study modern assessment practices, structural-functional analysis helped isolate the components of the learning outcome assessment system, and logical modeling was used to develop proposals for improving educational procedures. As a result of the study, it was established that the use of generative neural networks increases the risks of inauthenticity in completed tasks, reduces assessment transparency, and causes an information imbalance between participants in the educational process. A comprehensive set of solutions is proposed, covering clear regulation of digital service usage, transformation of control methods by combining oral and written formats, development of tasks emphasizing critical thinking, and the application of originality verification technologies. The study substantiates the expediency of transitioning to competency-based assessment focused on analysis, interpretation, and argumentation. The conclusions state that ensuring academic integrity in the context of generative AI requires a combination of regulatory frameworks, modernization of assessment procedures, and the development of the digital culture of educational process participants. Implementing the proposed solutions will contribute to increasing the objectivity of learning achievement assessment and strengthening trust in educational institutions.