Speaker
Description
Artificial intelligence reached thesis writing before universities had fully agreed on how it should be used. This paper looks at what students actually did with AI when preparing professional bachelor theses in one Lithuanian higher education institution. It compares two post-defence cohorts, from 2024 and 2025, and draws on 258 survey responses from a target population of 399 students. The questionnaire asked about thesis-related tasks, specific tools, perceived benefits and risks, attitudes towards AI, and open comments.
The results show a fairly clear pattern. Students did not use AI in every part of thesis writing, and they did not treat it as a replacement for their own work. They turned to it most often for translation, paraphrasing and summarising, grammar and style checking, and literature search. Much less often, it was used in areas more closely tied to academic ownership, such as topic definition or data analysis. Tool use was also concentrated rather than broad: ChatGPT and Google Scholar stood out clearly, while most other tools remained secondary. Students valued AI because it saved time and made the writing process easier to manage, but they also saw the risks. Their answers repeatedly pointed to inaccurate output, weaker originality, and the need to verify what AI produces instead of trusting it automatically. A positive relationship was also found between more frequent AI use and a more favourable view of its place in academic work.
One finding is especially clear. Students want more explicit guidance from lecturers on what counts as acceptable AI use in thesis writing. Taken together, the results suggest that AI has already become part of everyday thesis-writing practice, but mainly as a support tool. The practical question is no longer whether students use AI. The real question is how institutions define its place without weakening responsibility for the final text.