Undergraduate nursing students’ experiences of conducting graduation theses and using generative artificial intelligence: a qualitative study

Completing a graduation thesis is a sequential process that involves topic selection, proposal formulation, data collection and analysis, and thesis writing. Undergraduate nursing students generally encounter difficulties such as insufficient ability in literature retrieval and critical evaluation, weak statistical analysis skills, and limited academic writing experience during these stages. Moreover, the supervisor is unable to provide continuous personalized guidance throughout the entire process. The genAI offers new support possibilities, but currently there is a lack of qualitative evidence regarding how nursing students can experience and utilize genAI throughout the entire process. Therefore, this study explores the experiences of undergraduate nursing students in conducting graduation theses and using genAI. This study conducted semi-structured interviews with 19 undergraduate nursing students selected from a medical college in China from April 30th to May 31st, 2026. The subjects were recruited from the college and its affiliated hospital through convenience sampling, and the approval from the hospital’s ethical review was obtained. The data were processed using Colaizzi’s seven-step analysis method, supplemented by consensus meeting discussions. This study identified 38 themes across three dimensions: emotions, pain points, and cognitive and evaluative aspects of generative artificial intelligence-assisted writing. All 19 participants reported using genAI during the thesis process. During the process of completing their graduation thesis, undergraduate nursing students will experience periodic emotional fluctuations and multiple pain points. Generative AI is widely used and is considered helpful in academic writing, efficiency improvement, and framework organization, but students have also expressed concerns about generative AI generating false literature, language stiffness, and academic misconduct. Nursing educators should provide phased methodological and psychological support and formulate clear guidelines for the use of generative AI to regulate students’ behavior.