Exploring perceptions of data risks in AI-enabled nursing research

Abstract

Aim: To explore the data risk perception structure and connotation in the entire process of generative AI-enabled nursing research and to identify healthcare management and training needs as applied to digital health developments. Background: Nursing research highly relies on contextualized and unstructured data. General generative AI still faces shortcomings in professional adaptation, data governance, and responsibility definition, which may lead to risks such as privacy leaks, amplified bias, academic misconduct and accountability vacuums. The study focusses on the perceptions of Future nursing professionals. Methods: Purposeful maximum variance sampling was used to recruit 20 participants from 3 universities, and semi-structured one-on-one interviews were conducted. The report followed the COREQ Protocol checklist. Results: Five data risk awareness themes were identified: data adaptation risk, data security risk, data quality risk, data ethics risk, and response risk, presenting risk concerns throughout the entire process of “use—generation—sharing—responsibility”. Conclusion: The data risk perception of nursing master’s students regarding generative AI-enabled nursing research presents a clear five-dimensional structure, unfolding along the chain of “input—processing—output—diffusion—attribution.” This structure supports the development of a framework for defining boundaries of AI use, data governance, ethical compliance, and capacity building in nursing research settings and digital health developments.