Study Finds Students Prefer ChatGPT for Academic Work, Identifies Main Motivators

Researchers from the Faculty of Informatics at Eötvös Loránd University (ELTE) recently examined which elements influence university students’ intention to continue using ChatGPT as an educational tool. Using an interpretable machine‑learning model on survey responses, they pinpointed a small set of variables that most powerfully forecast ongoing adoption of the AI chatbot.

The researchers collected responses from a varied group of undergraduates spanning several fields, requesting them to evaluate ChatGPT on aspects like perceived usefulness, interaction ease, and trust in its outputs. The analytical framework, built to reveal the predictive model’s inner workings, singled out three key drivers: how relevant students found the tool for their coursework, their confidence in the accuracy of its answers, and the degree to which it facilitates personalized study routines.

Learners who indicated that ChatGPT assisted them in understanding difficult topics or in drafting assignments more quickly showed a significantly higher intention to keep using it. In contrast, worries about misinformation or excessive dependence on AI‑produced text lowered the probability of ongoing use. The study also detected a secondary effect of peer endorsement: students who saw peers benefiting from the chatbot were more inclined to intend its adoption.

The results emerge as higher education debates the place of generative AI. Numerous universities are trialling policies that embed AI into curricula, whereas others wrestle with academic‑integrity issues. By isolating the factors that fuel student enthusiasm, the ELTE research provides a data‑driven guide for educators aiming to employ AI responsibly—highlighting the need for transparent output validation and instruction that aligns the technology with teaching objectives.

Looking forward, the authors recommend that universities boost long‑term usage by incorporating AI‑literacy modules into orientation sessions and supplying faculty with guidelines for crafting AI‑compatible assignments. These steps could close the trust gap and lessen plagiarism risks, while still capitalising on the efficiency benefits students appreciate.

As AI further infiltrates academic settings, the ELTE study highlights the need to grasp user motivations. By matching technological potential with student requirements and institutional safeguards, higher education can more effectively steer through the shifting terrain of AI‑enhanced learning.