Demo Session: Rethinking Computer Assistance in Qualitative Research in the Age of Generative AI: The QDA Miner Approach

Computer assistance in qualitative research has long been designed to support, rather than replace, the interpretive role of the researcher. This tutorial examines generative AI from a critical perspective, identifying what current large language models can and cannot reliably contribute to qualitative analysis. Particular attention is given to limitations related to reproducibility, transparency, analytic control, sensitivity to research design, and the potential introduction of bias or consensual drifts. The tutorial will also discuss a frequently overlooked factor in generative AI use for qualitative research: the differences between Web-based interaction and API-based workflows and their methodological implications.

Rather than presenting generative AI as a substitute for existing CAQDAS practices, the session introduces strategies we have developed to address these limitations by combining GenAI with more traditional AI and qualitative research techniques. By emphasizing complementary and hybrid workflows, the tutorial aims to help qualitative researchers make informed, methodologically sound decisions about when and how AI can assist their analytic work, while preserving the central role of human interpretation.