https://learningletters.org/index.php/learn <p><em>Learning Letters </em>focuses on rapid publication of promising research in areas of learning analytics, educational technology, human and artificial cognition, artificial intelligence and education, learning design, and learning sciences. </p> en-US maarten.delaat@unisa.edu.au (Maarten de Laat) maarten.delaat@unisa.edu.au (Maarten de Laat) Thu, 17 Oct 2024 23:23:08 -0700 OJS 3.3.0.7 http://blogs.law.harvard.edu/tech/rss 60 https://learningletters.org/index.php/learn/article/view/35 <p>Concerns around Generative Artificial Intelligence (GenAI) in higher education have so far largely centred on assessment integrity, resulting in fundamental questions about students’ broader engagement with these tools remaining underexplored. This paper reports on the findings of a survey that forms part of a wider study, comprising the first empirical investigation of GenAI use by university students as a method of engaging with their academic readings. Our survey of 101 students shows that over half of all students surveyed used GenAI to some degree as a method of engaging with their unit readings. Our findings suggest that students turn to these tools in response to time constraints and conceptual difficulties, while maintaining complex attitudes toward their use: most welcome their availability, yet few report trusting or relying on GenAI-generated interpretations or summaries of texts. Importantly, our data reveals substantial demographic variations in usage patterns, with international students and those taking subjects as electives showing significantly higher rates of AI use to assist with reading. This suggests GenAI tools may be serving as important mediators for attempting to overcome epistemic barriers to learning, particularly for students who face additional linguistic or disciplinary challenges.</p> Thomas Corbin, Yifei Liang, Margaret Bearman, Tim Fawns, Gene Flenady, Paul Formosa, Lucinda McKnight, Jack Reynolds, Jack Walton Copyright (c) 2024 Learning Letters https://learningletters.org/index.php/learn/article/view/35 Mon, 09 Dec 2024 00:00:00 -0800 https://learningletters.org/index.php/learn/article/view/34 <p>This study employed a questionnaire survey using the California Critical Thinking Skills Test (CCTST) on a sample of 240 Chinese university students. The results obtained were then compared with 240 questionnaire results created by ChatGPT in order to investigate variations in critical thinking skills across different dimensions. The study uncovered that, although ChatGPT had exceptional performance in inference and deductive reasoning, it fell behind university students in the areas of evaluation and inductive reasoning. Meanwhile, with respect to analysis, the capabilities of both college students and ChatGPT were not significantly different. These findings provide essential insights for the advancement of artificial intelligence and human progress.</p> Shan Li, Min Fan, Gurnam Kaur Sidhu Copyright (c) 2024 Learning Letters https://learningletters.org/index.php/learn/article/view/34 Thu, 17 Oct 2024 00:00:00 -0700