Evaluating Human and LLM-Generated Thematic Analysis in HRI for Vulnerable Populations: A Comparative and Ethical Analysis

Alva Markelius; Fethiye Irmak Dogan; Julie Bailey; Hatice Gunes

cs.ROAug 14, 2026

Evaluating Human and LLM-Generated Thematic Analysis in HRI for Vulnerable Populations: A Comparative and Ethical Analysis

Authors: Alva Markelius, Fethiye Irmak Dogan, Julie Bailey, Hatice Gunes

Organizations: Affective Intelligence and Robotics (AFAR) Lab, Department of Computer Science and Technology, University of Cambridge, UK. · Faculty of Education, University of Cambridge, UK.

Abstract

Thematic analysis (TA) has long been regarded as an inherently human, reflexive, and interpretive process. However, the extent to which LLM-generated TA is appropriate for Human-Robot Interaction (HRI) research involving vulnerable populations remains largely unexamined and raises critical questions about validity and ethics, particularly in sensitive research contexts. This paper presents a comparative study of human- and LLM-generated TA in an HRI context with a focus on vulnerable populations. We evaluate both objective and semantic agreement between human- and LLMgenerated themes, and examine whether observed divergences reflect systematic interpretive patterns with ethical significance. Our analysis investigates whether LLM-generated TA risks marginalising or misrepresenting the experiences of vulnerable participants, with implications for researchers employing LLM-assisted TA in HRI.

Figures & tables

Fig. 1: Clustering consistency of human and LLM thematic analysis.

Fig. 2: Confusion matrix between human and LLM themes.

Fig. 3: Cosine similarity of human and LLM codes, subthemes, and themes averaged by human themes.

Utterance	Human Code	LLM Code	Cos sim.	IG	RF	BMR
S1. “I’m very aware of everything that we do, every every subject, including technology and sciences, are biased. Currently very, very biased so. I’m a bit worried that these robots might have certain. Unintentional agendas and viewpoints, like sexism, racism, and ableism.”	Concern over ableism and other biases	Concern about AI bias and discrimination	0.639	Medium	Medium	High
S2. “The hand gestures were quite lifelike while they look like actually a human hand gestures. So that was quite nice to sort of experience”	Body language	Lifelike gestures positively noticed	0.561	Medium	Medium	Low
S3. “It was kind of weird to use it’s hands to, like make up for it. And it was just. I’ve never seen something like move its hands so much and for someone who has anxiety that’s quite like you just heard people by their hand like they’re like actions in the way it just felt like very overwhelming.”	Anxiety could be exacerbated	Excessive hand movement overwhelming for anxious users	0.538	Low	Low	High
…
D3. “Now he’s like a conversational agent. It provides you this information and. He listened to you like as a sounding board. It would be great for him to actually help me with something, maybe help me write an essay.”	Sense of being listened to	Robot useful for practical tasks like emailing	0.085	Low	Low	High
D2. “At first, when I was interacting with robots because it was very lifelike so I figured it was like OK, just like like a normal conversation with a real life person. So I guess I got, I took the time to develop what I wanted to ask the question, but then the robot didn’t allow me to tell you to continue with the questions.”	Lifelike	Robot interrupts before user finishes speaking	0.065	Low	Low	Medium

TABLE I: Qualitative comparison and evaluation of codes. The top 3 similar (S1, S2, S3) and dissimilar (D3, D2, D1) codes are presented based on their cosine similarity. IG: Interpretive Grounding, RF: Representational Fidelity, BOR: Bias/Marginalisation Risk.

Human-Robot Interaction

Human-robot interaction studies how people and robots communicate, coordinate, and safely work or live together.

See all 453 papers

Designing an Auditable LLM-Supported Workflow for Qualitative Thematic Analysis
Aug 31, 2026Nadia Jul Jeldtoft, Tariq YousefLLM-Assisted Annotation
Department of Culture and Language 2Department of Mathematics and Computer Science University of Southern Denmark
TAMA: A Human-AI Collaborative Thematic Analysis Framework Using Multi-Agent LLMs for Clinical Interviews
Mar 26, 2025Huimin Xu, Seungjun Yi, Terence Lim +9Multi-Agent LLMs Clinical NLP
Department of Biomedical Engineering, University of Texas at Austin, USA · McCombs School of Business, University of Texas at Austin, USA · School of Information, University of Texas at Austin, Austin, TX, USA and Dell Medical School, University of Texas at Austin, Austin, TX, USA
Automated Thematic Analysis for Clinical Qualitative Data: Iterative Codebook Refinement with Full Provenance
Mar 9, 2026Seungjun Yi, Joakim Nguyen, Huimin Xu +9LLM-Assisted Annotation Iterative Refinement
Department of Biomedical Engineering, University of Texas at Austin, Austin, TX, USA · Department of Computer Science, University of Texas at Austin, Austin, TX, USA · School of Information, University of Texas at Austin, Austin, TX, USA +6

Aug 31, 2026J/K move · Enter open · S save

Designing an Auditable LLM-Supported Workflow for Qualitative Thematic Analysis

Nadia Jul Jeldtoft, Tariq Yousef

Department of Culture and Language 2Department of Mathematics and Computer Science University of Southern Denmark

LLM-Assisted Annotation

Open paper AskPDF arXiv

Large Language Models (LLMs) offer new possibilities for scaling qualitative analysis, but existing applications often provide limited methodological transparency regarding how qualitative methods are translated into computational procedures. This paper presents an auditable and privacy-preserving computational operationalization of inductive and latent Thematic Analysis (TA). This paper first derives five design principles from the methodological requirements of TA and the conditions introduced by LLM-based inference: preserving interpretative context, maintaining traceable relationships between empirical material and analytical outputs, representing analytical constructs and reasoning explicitly, constraining LLM inference to interpretative tasks, and enabling privacy-preserving local deployment. Second, it presents a proof-of-concept for a two-phase workflow that operationalizes these principles by combining interpretative LLM inference with deterministic procedural control to generate codes, analytical justifications, themes, and theme descriptions while preserving explicit links to the source material. Third, it proposes an evaluation framework combining structural comparison with human-led TA and independent expert assessment of analytical quality. The evaluation is conducted on semi-structured Danish interview transcripts. and the results shows that the workflow produces code-level outputs with coverage broadly comparable to human annotations and highly rated analytical justifications, while generating a more compressed thematic structure characterized by fewer and broader themes. The findings demonstrate the feasibility of auditable LLM-supported TA through a modular workflow designed to scale to larger datasets, accommodate different LLMs, and support transfer across research domains, with domain adaptation primarily requiring adjustments to the prompting strategy.

Abstract

Human-Robot Interaction

Explore similar work

Designing an Auditable LLM-Supported Workflow for Qualitative Thematic Analysis

TAMA: A Human-AI Collaborative Thematic Analysis Framework Using Multi-Agent LLMs for Clinical Interviews

Automated Thematic Analysis for Clinical Qualitative Data: Iterative Codebook Refinement with Full Provenance