Students Favor ChatGPT for Learning, New Study Reveals Key Drivers

A recent investigation by the Faculty of Informatics at Eötvös Loránd University (ELTE) has mapped the factors that shape whether university students intend to keep using ChatGPT as a learning aid. By applying an interpretable machine learning model to survey data, researchers identified a handful of variables that most strongly predict continued adoption of the AI chatbot.

The study surveyed a diverse cohort of undergraduates across multiple disciplines, asking participants to rate their experiences with ChatGPT on dimensions such as perceived usefulness, ease of interaction, and trust in the generated content. The analytical framework, designed to expose the inner logic of the predictive algorithm, highlighted three primary determinants: the perceived relevance of the tool to coursework, the degree of confidence in answer accuracy, and the extent to which the platform supports personalized study habits.

Students who reported that ChatGPT helped them grasp complex concepts or draft assignments more efficiently were markedly more likely to plan future use. Conversely, concerns about misinformation or over‑reliance on AI-generated text reduced the likelihood of continued engagement. The research also uncovered a secondary influence of peer endorsement; learners who observed classmates benefitting from the tool expressed higher intentions to adopt it themselves.

These findings arrive amid a broader debate within higher education about the role of generative AI. While many institutions are experimenting with policies that integrate AI into curricula, others grapple with academic integrity concerns. By pinpointing the elements that drive student enthusiasm, the ELTE study offers a data‑backed roadmap for educators seeking to harness AI responsibly—emphasizing transparent validation of outputs and training that aligns the technology with pedagogical goals.

Looking ahead, the authors suggest that universities could improve sustained usage by embedding AI literacy modules into orientation programs and by providing faculty with guidelines for designing AI‑compatible assignments. Such measures could address the trust gap and mitigate the risk of plagiarism, while still leveraging the efficiency gains that students value.

As AI continues to permeate academic environments, the ELTE research underscores the importance of understanding user motivations. By aligning technological capabilities with student needs and institutional safeguards, higher education can better navigate the evolving landscape of AI‑enhanced learning.