Juraj Fabuš
2025 AI and higher education

What ChatGPT can actually do in management education

Original title: The Use of AI Tools in Managerial Education

Juraj Fabus, Miriam Garbarova, Lukas Vartiak, Ludmila Mitkova (2025)

ChatGPT did not arrive in university teaching through policy documents but through students. So with my co-authors I tested it directly on the tasks students of Electronic Commerce and Management deal with every day — essay writing, comprehension and problem-solving.

Why it is worth reading

Instead of speculating about whether AI will "transform education", it shows concretely where it helps a student and where it leaves them stranded.

What to take away

  • ChatGPT strengthens student autonomy and critical thinking — provided the student treats it as something to argue with rather than an answer machine.
  • It runs into two hard limits: no reliable access to up-to-date data, and no source citations. For academic work both matter a great deal.
  • So the question is not "allow or ban" but which kinds of assignments it suits and which it does not.

The starting point

The debate about artificial intelligence at universities circled for a long time around whether to allow it. But reality has already answered that one — students use it either way.

It is more useful to know what it is good for and what it is not. So my co-authors and I decided to stop theorising and try it.

What we did

We focused on teaching in the Electronic Commerce and Management study programme, combining qualitative content analysis with experimental testing on selected students.

We tested three kinds of task that make up a substantial part of their ordinary work: essay writing, comprehension, and problem-solving.

What came out of it

The good news: ChatGPT strengthens student autonomy and critical thinking. Not because it thinks for them, but because it gives them something to disagree with. A student looking at a draft answer approaches the topic differently from one staring at a blank page.

The bad news is specific and comes down to two points. The model has no reliable access to current data, and it does not cite sources. For academic writing that has to rest on verifiable claims, both are serious limitations.

What this means for teaching

The conclusion we reached is neither “ban it” nor “let it be”. It is a distinction: for assignments about understanding and argument, AI is a useful aid. For assignments resting on current data and citation, the model is more likely to harm a student than help them.

From a teacher’s point of view that means less work policing students and more work on how the assignment is built in the first place.

Bibliographic details

Authors
Juraj Fabus, Miriam Garbarova, Lukas Vartiak, Ludmila Mitkova
Published in
Management Theory and Studies for Rural Business and Infrastructure Development
Year
2025
Volume
47
Issue
2
Pages
251-259
Publisher
Vytautas Magnus University
DOI
10.15544/mts.2025.19
Keywords
Artificial Intelligence (AI), ChatGPT, Managerial Education, Generative Models, Academic Writing
Show abstract

Official abstract in its original wording.

This paper investigates the use of AI tools—particularly ChatGPT—in managerial education, focusing on its application in the study of E-Commerce and Management. The paper employs qualitative content analysis and experimental testing with selected university students to evaluate the effectiveness of ChatGPT in learning tasks such as essay writing, comprehension, and problem-solving. The findings demonstrate that while ChatGPT enhances student autonomy and critical thinking, it also presents challenges, such as limited access to up-to-date data and the absence of source citations. The paper concludes with practical recommendations for responsibly integrating AI into educational processes to support academic achievement and curriculum innovation.

Cite

Fabus, J., Garbarova, M., Vartiak, L., & Mitkova, L. (2025). The Use of AI Tools in Managerial Education. Management Theory and Studies for Rural Business and Infrastructure Development, 47(2), 251-259. https://doi.org/10.15544/mts.2025.19

Download RIS

Back to research