GENAI В АКАДЕМІЧНОМУ ПИСЬМІ: ЕМПІРИЧНЕ ДОСЛІДЖЕННЯ РЕКОНСТРУКЦІЇ КОНЦЕПТУАЛЬНИХ КОМПОНЕНТІВ НАУКОВОГО ДОСЛІДЖЕННЯ

The article addresses the topical issue of using generative artificial intelligence (GenAI) in working with academic texts by combining practical work with GenAI, the analysis of dissertation abstracts, and the development of students’ critical attitudes toward AI-generated information. This study investigates the ability of generative artificial intelligence to reconstruct selected conceptual components of academic discourse without direct access to the corresponding dissertation abstract. The study analyzes contemporary scholarly approaches to the use of generative artificial intelligence in academic discourse. The aim of the study is to empirically determine the extent to which GenAI can reproduce the relevance, research aim, and research focus of an academic study based on a minimal set of input data. The methodology involved an experimental procedure with 25 undergraduate students enrolled in the program “Language and Literature (German). Translation” within the field of Germanic languages and literatures. The empirical material comprised 25 abstracts of successfully defended dissertations on Germanic linguistics. Two generative AI systems, ChatGPT and Gemini, were used in the experiment. Following a standardized prompting procedure, the systems were provided with the research topic, but the corresponding dissertation abstract was not made available to them. The AI-generated responses were compared with the corresponding conceptual components of the dissertation abstracts using comparative content analysis. The results indicate that GenAI can semantically reconstruct selected characteristics of an existing academic study, including its relevance, research aim, and focus. At the same time, the findings do not indicate textual copying of dissertation abstracts but rather suggest the possibility of reconstructing their content-related characteristics from limited input data. These results indicate the considerable potential of GenAI as a tool for supporting academic writing. Practical implications. The findings may be used to develop university curricula that establish a balanced approach to integrating generative artificial intelligence into the educational process while preserving fundamental principles of academic integrity and values.