DeTAILS: Deep Thematic Analysis with Iterative LLM Support

AI can accelerate qualitative research while preserving human judgment: DeTAILS demonstrated that LLM assistance reduced researcher workload in thematic analysis—a time-intensive qualitative method—while maintaining strong alignment between AI-generated codes and researcher refinements, suggesting AI tools can scale interpretive work without replacing expert decision-making.

Transparency and user control are essential for researcher trust in AI tools: Participants reported that trust in DeTAILS grew through transparency about how the system worked and their ability to review, modify, and reject AI outputs, highlighting that responsible AI design for research requires giving users meaningful oversight rather than treating AI suggestions as authoritative.

Interactive feedback loops help researchers critically engage with AI rather than passively accept it: The study found that DeTAILS's iterative design—where researchers reviewed and refined AI-generated clusters and themes—prompted "reflexive engagement" with outputs, meaning researchers actively questioned and improved upon AI suggestions rather than uncritically adopting them.

Human-AI collaboration in research requires preserving "analytic agency": The toolkit was specifically designed to maintain researchers' interpretive authority throughout the analysis process, suggesting that responsible AI integration in knowledge work should enhance rather than automate human expertise and decision-making.

Feasibility alone isn't enough—researcher experience matters for adoption: Beyond showing the system worked, the study measured perceived usefulness and workload reduction among 18 researchers, demonstrating that responsible AI tools must be evaluated not just on technical performance but on whether they actually improve researchers' workflows and experiences.