University students' purposeful engagement with generative AI: differentiating tools, functions and goals
Julio Gimenez, Katherine Mansfield, Richard PatersonPurpose
This paper aims to examine university students' choice of ChatGPT, Grammarly and Gemini for academic tasks. It contributes to debates on artificial intelligence (AI) in higher education by analysing students' differentiated engagement with generative artificial intelligence (GenAI) tools and proposing guidelines for GenAI literacy (GenAIL) development.
Design/methodology/approach
The study used a mixed-methods online survey combining Likert-scale, multiple-choice and open-ended items. In total, 441 students completed the survey. Quantitative data were analysed using a chi-square test and qualitative data through descriptive thematic categorisation.
Findings
The paper provides empirical insights into the relationship between AI tool choice and academic task. ChatGPT was strongly linked to idea generation and planning, Grammarly to writing and feedback, while Google Gemini appeared comparatively diffuse across task categories.
Research limitations/implications
The contextualised sample limits generalisability, and the study would have been strengthened by interviews or focus groups. Future research could adopt longitudinal mixed-methods designs.
Practical implications
The paper highlights the need to embed critical digital literacy development to support purposeful and informed GenAI use in higher education and provides guidelines for implementation.
Originality/value
By examining the statistical association between academic tasks and tool selection, the paper contributes to emerging understandings of how students choose specific AI tools and proposes guidelines for GenAIL development.