AI in Qualitative Research | audiotranskription

Since 2024, we at audiotranskription have been investigating how generative AI can be used in qualitative research without delegating interpretation, verification, and responsibility to a language model.

Our work at the intersection of transcription, qualitative analysis, software development, and the teaching of research methods has resulted in procedures, prompt templates, practical guides, workshop reports, and scholarly articles.

This section brings together these materials and outlines the methodological and data protection-related conditions under which we consider the use of AI in research projects to be justifiable.

Procedures & Resources

Hybrid interpretation and serendipity prompting integrate LLMs as limited catalysts into a research-driven analysis process.

For both methods, you’ll find methodological justifications, prompt templates, instructions, and documented sample analyses, as well as guidance on transparent documentation.

About Procedures & Resources

Presentations & Workshops

Webinars, presentations, and workshops provide an opportunity to discuss the methods, their prerequisites, and their limitations using concrete examples—either online or at your location.

Workshops & Presentations

Publications

Our workshop reports and papers document what the use of AI in qualitative research actually achieves and where its limitations lie—for example, in terms of the effort required for validation, local models, or commercial AI infrastructures.

Also: the lecture notes as an overview and all publications for further reading and citation.

See Publications

Data Protection & Data Sovereignty

Qualitative interview data often remains personally identifiable despite anonymization.

The issue here is which prerequisites must be clarified before AI processing can take place (consent, an ethical review if necessary, and responsibilities) and which technical processing methods are even feasible under these conditions.