Exploring the Use of AI in Qualitative Analysis: A Comparative Study of Guaranteed Income Data
Leah Hamilton, Desha Elliott, A. J. Quick, Simone G.V.S. Smith, Victoria Choplin
Abstract
This study explores the potential of the AI chatbot ChatGPT to supplement human-centered tasks such as qualitative research analysis. The study compares the emergent themes in human and AI-generated qualitative analyses of interviews with guaranteed income pilot recipients. The results reveal that there are similarities and differences between human and AI-generated analyses, with the human coders recognizing some themes that ChatGPT did not and vice versa. The study concludes that AI like ChatGPT provides a powerful tool to supplement complex human-centered tasks, and predicts that such tools will become an additional tool to facilitate research tasks. Future research could explore feeding raw interview transcripts into ChatGPT and incorporating AI-generated themes into triangulation discussions to help identify oversights, alternative frames, and personal biases.