Tailored behavioral personas of large language model use among Chinese nursing students: A qualitative study.
TL;DR
Multiple factors shape nursing students' diverse patterns of interaction with large language models, and future interventions should be tailored to these specific personas to mitigate the risks of cognitive outsourcing and foster the responsible integration of large language models into nursing education.
Abstract
Background
With the integration of large language models into nursing education, nursing students' patterns of use have attracted growing attention and concern. While existing research focuses on determinants such as attitudes and intentions, it lacks insight into their actual patterns of use.
Objective
To identify the tailored behavioral personas of nursing students regarding large language models and explore the characteristics of interaction patterns within different personas, thereby clarifying the potential risks associated with these patterns of use.
Methods
From October to December 2025, a descriptive qualitative study was conducted involving 22 nursing students in China via semi-structured interviews. Purposive sampling with a maximum variation strategy was employed to select nursing students. Data were analyzed using content analysis. Through the extraction of interaction tags and behavioral dimensions, user personas were constructed to characterize students' patterns of use.
Results
Five key dimensions of students' interactions with large language models were extracted: cognitive relationships, interaction strategies, verification strategies, psychological experiences, and risk perceptions. Based on these dimensions, four user personas were identified: the efficiency-quality trade-off persona, the capability-compensating persona, the prudent-assistance persona, and the cognitive outsourcing persona.
Conclusion
Multiple factors shape nursing students' diverse patterns of interaction with large language models. Future interventions should be tailored to these specific personas, combining critical thinking training with technical and ethical support to mitigate the risks of cognitive outsourcing and foster the responsible integration of large language models into nursing education.