"Where No One Has Gone Before": Questions to Ensure the Ethical, Rigorous, and Thoughtful Application of Artificial Intelligence in the Analysis of HIV Research.

ChatGPT, an artificial intelligence (AI) system released by OpenAI on November 30th, 2022, has upended scientific and educational paradigms, reshaping the way that we think about teaching, writing, and now research. Since that time, qualitative data analytic software programs such as ATLAS.ti have quickly incorporated AI into their programs to assist with or even replace human coding. Qualitative research is key to understanding the complexity and nuance of HIV-related behaviors, through descriptive and historical textual research, as well as the lived experiences of people with HIV. This commentary weighs the pros and cons of the use of AI coding in HIV-related qualitative research. We pose guiding questions that may help researchers evaluate the application and scope of AI in qualitative research as determined by the research question, underlying epistemology, and goal(s). Qualitative data encompasses a variety of media, methodologies, and styles that exist on a spectrum underpinned by epistemology. The research question and the data sources are informed by the researcher's epistemological viewpoint. Given the heterogeneous applications of qualitative research in nursing, medicine, and public health there are circumstances where qualitative AI coding is appropriate, but this should be congruent with the aims and underlying epistemology of the research.

Duke Scholars

Author Michael Vernon Relf School of Nursing