From Support to Dependency: Exploring Student Perceptions of Generative AI

ABSTRACTThe rapid integration of Generative Artificial Intelligence (GenAI) into education highlights the need for empirical insights into student perspectives to inform institutional policies, pedagogical strategies, and social practices. This study explored Italian university students’ perceptions of GenAI’s benefits and risks across personal, social, and academic domains. A total of 1,347 students from 24 Italian universities completed an online questionnaire. This manuscript focuses on two open-ended questions investigating perceived advantages and disadvantages of ChatGPT and similar systems, analyzed through manual coding using MaxQDA software. Results show students predominantly adopt an instrumental view of GenAI: operational benefits made up 41.6% of responses, followed by informational (14.3%) and educational advantages (10.9%). The most cited disadvantages were cognitive and learning risks (29.6%), problematic behaviors (23.6%), and technical limitations (15.4%). These findings reveal an ambivalent stance: students value GenAI’s efficiency and learning support but express concerns about intellectual dependency, diminished critical thinking, and threats to academic integrity. This study makes several key contributions: it provides the first large-scale qualitative investigation (N=1,347) of GenAI perceptions in Italian higher education, bridging the gap between qualitative depth and quantitative generalizability; it offers culturally situated insights from a non-Anglophone context, enriching the predominantly Anglo-centric literature; and it delivers evidence-based recommendations for institutional policy development grounded in actual student experiences rather than theoretical assumptions. Such insights are crucial for understanding the views of a key stakeholder group in higher education and for guiding the development of ethical usage policies, structured teacher training, and tailored student orientation initiatives. These measures aim to promote technological governance skills that preserve intellectual autonomy while harnessing GenAI’s operational benefits.