Artificial Intelligence in Qualitative Research: Methodological Turn

This study explores the methodological challenges that arise when employing generative AI (GAI) in qualitative research. ChatGPT is introduced as a collaborator in the research process, working alongside two human researchers – authors of this paper. This partnership serves as a tool for qualitative data analysis and ensures data triangulation.
Research questions. How can the interpretability of AI-generated insights be ensured to maintain the trustworthiness of the research findings in qualitative research, particularly when applied to narrative data from TedTalks lectures? In what ways can GAI enhance traditional qualitative research methodologies in terms of efficiency and depth of analysis?
The primary goal of this study was to examine the integration of ChatGPT into the qualitative research process. To investigate the topic, we have chosen the case of the Situational Analysis methodology, developed by Adele Clarke. ChatGPT was employed as a collaborative partner in data collection from TedTalks and subsequent data analysis.
The study yields valuable results in the form of situational and social world-arena maps, which are generated using ChatGPT's analytical capabilities. These maps provide insights into the complex interplay of ideas and concepts within the TedTalks .
Artificial Intelligence, when integrated into the qualitative research process, presents promising opportunities for data collection and analysis. The collaborative efforts of human researchers and GAI could enhance the rigor and depth of qualitative research. However, this integration also raises major debates and tensions, including concerns about the interpretability of AI-generated insights and ethical considerations regarding the role of GAI in research. The limitation on the number of tokens in ChatGPT leads to truncation of data and lack of context. Specific bias may be related to ChatGPT's training data that has a notable impact on the comprehensiveness and depth, also objectivity and impartiality, precision, and accuracy of the data analysis.