Generative AI and Students’ Academic Writing Practices: A Cross-Sectional Study of Irish Technological Universities

The swift integration of large language models in universities has ignited debate regarding their influence on student writing and creative development. Most studies focus on efficiency and task completion, with limited examination of the lasting cognitive and innovative effects of repeated algorithmic assistance. Addressing this gap, this paper utilizes a quantitative survey of 180 undergraduate and postgraduate students across Munster Technological University, Technological University of the Shannon: Midlands Midwest, and Southeast Technological University. Stratification was performed by institution and field of study, including 61 female students (33.9%) and 119 male students (66.1%). The survey explored generative AI usage trends, perceptions of creative self-efficacy, and student reports of ideational similarity and originality. Results show ongoing tension: respondents favored ChatGPT for time-saving, idea generation, and clarification of complex concepts. However, many reported diminished trust in independent thought, reduced intellectual property, and growing concern that their work increasingly resembled that of peers using the same tools. Notably, 69.4% felt their ideas aligned more closely with other ChatGPT users, while 71.7% believed seemingly original ideas tended to mirror AI-generated suggestions. These findings suggest that while AI-assisted writing offers pedagogical benefits, it also brings anxieties surrounding cognitive dependence, standardization, and loss of intellectual distinctiveness, highlighting the need to protect independent judgment and disciplinary originality in AI-driven learning.