Interviews deliver detailed insights into human experiences and are a valued instrument for researchers across various disciplines. Conducting interviews is, however, notoriously resource-intensive. Large Language Models (LLMs) promise to support all stages of the interview process from developing interview protocols, acting as the interviewer, and analysing the collected data by coding utterances, extracting themes and finding quotes relevant to a research question. This session will explore the potential of LLMs for supporting the interview process.

Participants of this session will gain insights into the current state of research, applications, and future directions of using large models in research interviews. The session will include a presentation, providing an initial understanding of how LLMs can be integrated into qualitative research practice and a hands-on activity, using the AInterviewer tool developed to support interview practices in large-scale study settings.

Program

Introduction (5 minutes)

Talk: AInterviewer (10 minutes)
by Hjalmar Bang Carlsen, Associate Professor, SODAS, University of Copenhagen

Q&A (5 minutes)

Hands-on activity using AInterviewer (50 minutes)

Questions and discussion (20 minutes)

Organisers

• Tor-Salve Dalsgaard, Postdoc, University of Copenhagen
• Olga Larygina, PhD student, IT University of Copenhagen
• Aske Mottelson, Associate Professor, IT University of Copenhagen

Thanks to Willem van der Maden, Postdoc, IT University of Copenhagen

Level

Introductory – The participants should bring an understanding of the capabilities of LLMs or the challenges of conducting interview studies.