Abstract

Generative artificial intelligence (GenAI) is reshaping academic practice in higher education, yet evidence on how students evaluate its acceptable uses remains limited. This study surveyed 988 post-secondary students across 39 countries, recruited through the Prolific online research platform, on the acceptability of GenAI use across 22 academic tasks using a binary Yes/No format. Using an interpretive framework grounded in normative boundary-setting, academic integrity values, and academic socialization, the analysis reveals differentiated acceptability patterns: Students endorse GenAI for instrumental and research-support tasks while maintaining normative boundaries around evaluative and judgement-intensive activities. Gender differences are minimal; level of study is the primary source of variation, with graduate students reporting higher acceptability for research- and professionally oriented uses. Respondents residing in Canada (n = 58) reported significantly lower acceptance of four text-focused tasks, including academic composition and summarization, than the non-Canadian cohort. Findings have implications for institutional policy, pedagogy, and academic integrity frameworks.