Abstract

This exploratory study investigates how researchers across various academic disciplines utilize generative artificial intelligence (AI) tools in the scholarly writing process and what factors influence their use. Grounded in the Task-based Information Behavior theory and the CRediT (Contributor Roles Taxonomy) framework, this study employed a structured survey targeting graduate-level and professional researchers. A total of 113 valid responses were collected via an online questionnaire. The results show that approximately 90% of respondents had experience using generative AI tools when writing papers, with the highest usage reported in tasks related to “Writing - Review & Editing”, “Writing - Original Draft Preparation”, and “Conceptualization”. Differences in AI usage patterns were observed across academic fields and career stages in some tasks. Key challenges cited by respondents included concerns over the reliability and subject-matter competence of AI-generated outputs. In order to utilize generative AI responsibly, academic libraries need to provide practical support such as supporting access to paid tools, providing ethical usage guidelines, and curated comparative evaluations of different AI platforms. The purpose of this study is to provide basic data for this.