Ethical use and challenges of GenAI in academic writing: a study on university graduates in Bangladesh

Purpose This study aims to examine the pattern of generative artificial intelligence (GenAI) use among non-science backgrounds in academic writing, highlighting challenges and ethical issues from both students' and faculty members' perspectives. Design/methodology/approach This research adopts a predominantly qualitative approach, which is supplemented by limited quantitative elements. About 140 students from seven distinct social science disciplines contributed to data collection through observations. Additionally, 14 faculty members participated in in-depth interviews (IDI). Informed consent was obtained from all participants, and ethical considerations were strictly adhered to. Findings Students use various GenAI tools for conceptual understanding, idea generation, language enhancement, structural development, and formatting assistance. However, it creates overdependence among students and lessens their intellectual skills. They make mistakes in citations and utilize fabricated references, which are considered unethical. Research limitations/implications To ensure the best use of GenAI in academic writing, developing national and institutional policies is fundamental. Besides, introducing AI literacy, providing language support programs, strengthening analytical skills, and establishing ethical training modules on AI use might help remove barriers and increase students' competence. Originality/value Existing studies focus primarily on STEM students. In contrast, the unique GenAI challenges faced by social science students in countries like Bangladesh remain largely unexplored, and this major research gap is addressed.