The 31st ACM International Conference on Intelligent User Interfaces (IUI), Paphos, Cyprus

IUI 2026 Tutorial on AI for Qualitative Research:

A Comprehensive Field Guide to LLM-Supported Qualitative Research

Mon, 23 Mar, 2026 | 2:00 PM - 3:30 PM | Part One | Room 7

Mon, 23 Mar, 2026 | 4:00 PM - 5:30 PM | Part Two | Room 7

Due to the current travel uncertainties, ACM IUI 2026 has decided to reschedule the conference to July 2026. This tutorial has also been rescheduled accordingly.

Please refer to the ACM IUI 2026 website for the latest updates. Additional information will be shared soon. Thank you for your support.

We are closely monitoring the relevant travel advisories and will provide updates as soon as any new information becomes available.

Welcome

Welcome to our tutorial site for the AI4Qual tutorial at IUI 2026 in Paphos, Cyprus.

If you are going to attend this tutorial in person, please let us know at -> https://forms.gle/nXA8oP1TacxS1T8H6

Background of this Tutorial

Qualitative Research, with its unique ability to provide in-depth understanding of phenomena, capture subtleties, and uncover underlying mechanisms, holds an irreplaceable position across various fields such as Human-Computer Interaction (HCI), humanities, social sciences, and broader user research. It serves as a significant part for understanding human experience, social interactions, and cultural contexts.

With the technological advancements represented by Large Language Models (LLMs), an increasing number of people are starting to consider and explore the application of LLM-based tools in qualitative research. However, whether due to technologists not fully understanding qualitative research, or qualitative researchers lacking experience in deploying LLMs, or remaining concerns about applying LLMs in this field, this tutorial will help people better understand and correctly use the relevant technology in this domain.

Through this tutorial, we aim to address the following inherent and significant challenges in qualitative research:

Labor-Intensive: Every step, from designing interview scripts and conducting semi-structured interviews to transcription, and finally coding and thematic analysis, requires a substantial investment of manual time and effort.
Scalability Issues: Traditional qualitative methods struggle to efficiently process hundreds or even thousands of interview transcripts or field notes. Qualitative researchers often face a difficult trade-off between "depth" and "breadth."
Challenges in Knowledge Transfer and Consistency: The process of coding and thematic analysis heavily relies on the researcher's experience, making it challenging to maintain coding consistency and transparency across teams or projects, in particular for novices.

This tutorial is designed to help researchers of different skill levels understand and effectively utilize LLMs and related technologies and tools to facilitate, simplify, and empower their qualitative research endeavors.

Agenda and Tutorial Content Overview

This tutorial will guide you step-by-step through the two core application areas of LLMs in the qualitative research lifecycle.

Welcome Keynote

Part One: Application of LLM/MLLM in Data Collection: Enhancing Semi-Structured Interviews (≈ 60 mins)

Invited Keynote Speech for Part One

Active 1 - Teach: How LLMs/MLLMs Support Interview Design and Interviewer Preparedness

Learning Goals

Understand how LLMs can help refine research questions.
Learn to generate interview protocols with probes and follow-ups.
Explore ways to use MLLMs for semi-structured interview (collaborative modes and methods).
Discuss ethical and interaction considerations (hallucinations, bias, confidentiality, over-reliance on AI).

Content

Demonstration of prompt patterns for generating semi-structured interview guides.
Show how applying LLMs can facilitate interviews and support learning how to conduct semi-structured interviews.
Show how to use AI to adapt and adjust tone, cultural sensitivity, and domain specificity.
A demo of automated semi-structured interviews powered by AI.

Active 2 - Practice: Hands-On Interview Guide Creation and AI-Assisted Mock Interview Activity

Activity

Participants form rapid three-person research teams to practice LLM-enabled semi-structured interviewing.
They use an LLM/MLLM to:
- Generate/refine a semi-structured interview protocol for a chosen topic.
- Create follow-up questions tailored to different types of potential participants.
Participants run a short “mock interview” where:
- One person acts as the interviewer, one as the co-researcher (co-interviewer), and one as the participant (interviewee).
Debrief: What was useful? What required human judgment? Where did AI fail?

Outcome

A ready-to-use interview guide.
Practical understanding of the strengths and limits of LLM-generated interview interactions.
Methods for using LLMs to support novice interview learning and training.

Discussion and Reflection of Part One (≈ 15 mins)

Break (30 mins)

Part Two: Application of LLM in Qualitative Analysis: Coding and Thematic Analysis (≈ 60 mins)

Invited Keynote Speech for Part Two

Active 1 - Teach: Using LLMs for First-Cycle Coding and Analytical Memoing

Learning Goals

Understand how LLMs can assist with descriptive, in vivo, and process coding.
Learn to structure prompts for transparent, auditable coding workflows.
Explore strategies to maintain analytic rigor and researcher reflexivity.
Discuss risks: model bias, over-generalization, losing contextual nuance.

Content

Demonstration of coding a short transcript with and without the LLM.
Explain chain-of-thought prompting for qualitative reasoning.
Show how to generate memo notes, researcher reflections, and alternative interpretations.

Active 2 - Practice: Coding a Sample Transcript + Comparing Human vs. AI Interpretations

Activity

Participants receive a short qualitative transcript (provided by the workshop).
They perform:
- Manual first-cycle coding (5 mins).
- AI-assisted coding using a structured prompt (5–7 mins).
Small-group comparison:
- What codes overlapped? What diverged? Why?
- Did the AI introduce novel insights or impose unwarranted structure?
Use the AI to generate:
- Thematic clusters.
- A brief analytic memo.
Reflect on conditions under which AI helps vs. hinders interpretive work.