Profiles of Postgraduate Students’ Use of Artificial Intelligence in Writing: Antecedents and Outcomes

• In Event: Student Learning, AI, and Digital Tools

Abstract

Artificial intelligence (AI) is widely used in academic writing, yet little is known about how different AI-use patterns relate to learning outcomes. We adopted a person-centered approach to examine AI-use profiles, their predictors, and associated outcomes. Survey data from 1,073 UK postgraduate students were analyzed using latent profile analysis, revealing three profiles: transformation explorers, tech-driven users, and balanced users. Profile membership was predicted by academic discipline and AI-use frequency. Learning outcomes varied across profiles: transformation explorers reported the highest critical thinking and perceived improvement, tech-driven users showed the greatest motivation, while balanced users had moderate outcomes. By adopting a person-centered approach, this study pushes theoretical discussions beyond a binary view of AI as either “beneficial” or “detrimental”.

Authors

• Fangzhou Jin, The University of Hong Kong

• Chun Lai, The University of Hong Kong

• Chin-Hsi Lin, University of Hong Kong