Abstract:

Generative artificial intelligence (GenAI) tools such as ChatGPT are increasingly used by university students in their everyday academic work. Students often report that these tools help them understand course content, organize information, and improve the quality of their written assignments. However, an important question remains open: do these positive perceptions actually translate into better academic results? This study examines the relationship between students’ perceptions of generative AI and their actual academic performance. A mixed-methods research design was used, combining survey data with an analysis of students’ grades. The first stage of the study consisted of an anonymous survey conducted among university students in Spain. In total, 236 valid responses were collected. The questionnaire explored how frequently students use generative AI tools, what they use them for, and how they perceive their impact on learning and academic performance. The second stage of the study involved a comparative analysis of final project grades from the course Project Management across two academic years: 2022 and 2024. The year 2022 was treated as a reference point representing the period before the widespread use of generative AI in student work, while 2024 reflected a period in which such tools were already widely available and commonly used. The analysis included 58 final projects from 2022 and 67 projects from 2024, all evaluated using the same assessment rubric. The survey results show that generative AI has become a regular part of students’ study practices. All respondents reported using such tools, and most of them declared that they use them frequently. Students mainly rely on generative AI to search for information, improve grammar and writing style, generate ideas, and better organize their academic work. The majority of respondents believe that AI helps them understand course topics and improves their academic performance. At the same time, students are aware of some risks, such as excessive dependence on AI or difficulties in developing their own ideas. The analysis of students’ grades shows a modest improvement in academic results between the two analyzed years. The average project grade increased from 6.8 in 2022 to 7.5 in 2024. At the same time, the minimum grade increased from 1.0 to 5.0, while the maximum grade remained the same. A qualitative review of instructor feedback suggests that projects written in 2024 were often better organized, clearer in structure, and more complete in terms of formal requirements. However, improvements in deeper analytical thinking and interpretation of results were much less visible.

Overall, the findings suggest that students clearly perceive generative AI as a useful learning support tool, but its impact on measurable learning outcomes remains limited. Generative AI appears to help students improve the organization and presentation of their work, but it does not automatically lead to stronger analytical skills or deeper understanding. These results highlight the importance of designing learning tasks and assessment criteria that encourage critical thinking, interpretation, and independent reasoning when generative AI tools are used in higher education.

Keywords:

Generative artificial intelligence, higher education, student perceptions, academic performance.