The Role of Generative AI in the Modern University: Mapping Opportunities, Limits, and Risks for Academic Integrity

Abstract

In recent years, neural networks have demonstrated rapid growth and incredibly active development and improvement, being actively implemented in all spheres of human life, including education, which requires new research on their effectiveness for didactic purposes. The purpose of the article is to investigate the opportunities and threats from the use of generative AI in higher education, in particular, in the formation of professional English language training for students of natural sciences. In the course of this study, theoretical general logical (analysis, synthesis, generalization, induction, deduction) and empirical (pedagogical observation, pedagogical experiment, description, survey) methods were used. The theoretical and methodological basis is formed by the key provisions of cognitive psychology, cognitive pedagogy and AI theory. The article provides a critical analysis of the literature, discusses the advantages and disadvantages of implementing generative AI in the educational process; describes the didactic possibilities of using neural networks ChatGPT, Copilot, and AI Tutor; conducts a pedagogical experiment and describes its results on the use of neural networks in the process of learning academic English; develops methodological recommendations for the use of AI-based programs in the educational environment of higher education. It has been established that AI can take on different roles in learning English for academic purposes, so in the pedagogical experiment, students were allowed to use AI in the following three roles: teacher, editor, and foreign-language interlocutor. The study identified and examined the following pedagogical capabilities of ChatGPT for the development of students' cognitive and linguistic activity: criticality, creativity, cognitive interest, motivation to learn, reflexivity, and skill development in learning English. The results of the study showed the effectiveness of using AI to compile thematic glossaries, search for contexts of term use, test vocabulary, and conduct dialogues, which stimulated their cognitive interest. The results of the final tests showed higher achievements of the experimental group in geological English terminology, comprehension of scientific texts, and listening.