Advancing qualitative research through AI-powered interview innovations
Step Further Towards Automated Social Science: An AI-Powered Interview Platform
This study introduces MimiTalk, an AI-powered automated interview platform capable of conducting in-depth interviews autonomously while maintaining high data quality standards. Through evaluation involving 20 participants, we demonstrate significant improvements in interview efficiency while maintaining response quality comparable to traditional methods. Results show an average interview duration of 141.632 seconds with strong semantic coherence and comprehensive topic coverage.
Fengming Liu, Shubin Yu · 2025Research by Qualitati
Efficacy of Large Language Models in Short-Text Topic Modeling
This study explores the efficacy of large language models (LLMs) in short-text topic modeling, comparing their performance with human evaluation and Latent Dirichlet Allocation (LDA). In Study 1, we analyzed a dataset on chatbot anthropomorphism using human evaluation, LDA, and two LLMs (GPT-4 and Claude). Results showed that LLMs produced topic classifications similar to human analysis, outperforming LDA for short texts. In Study 2, we investigated the impact of sample size and LLM choice on topic modeling consistency using a COVID-19 vaccine hesitancy dataset. Findings revealed high consistency (80-90%) across various sample sizes, with even a 5% sample achieving 90% consistency.
Shubin Yu · 2024
Generative AI Policies in Academic Research: A Global Analysis of Top Universities
This study systematically analyzes Generative AI policies from 61 top global universities using large language models and manual validation. The research identifies 10 key policy themes including data privacy, transparency, and academic integrity, while revealing gaps in environmental sustainability and regular policy reviews. The findings provide a framework for responsible AI use in research and offer empirical evidence for policy standardization.
Renbin Song, Shubin Yu · 2025
The Comparison of AI Interviewers and Human Interviewers: A Large Scale Field Experiment
This study investigates the comparative effectiveness of AI and human interviewers through the development and deployment of two AI interview platforms—MimiTalk (online) and MimiCall (telephone-based). Drawing from Social Desirability Bias Theory, Emotional Contagion Theory, and Information Theory, we examine differences in questioning strategies, response patterns, and information quality.
MimiTalk Research Team · 2025
Comparative Analysis of Multi-Modal AI Interview Systems
A systematic comparison of different modalities in AI-driven interview systems, evaluating four distinct configurations: text-to-text, voice-to-text, video-to-text, and voice-to-voice interactions. The study assesses these modalities across key metrics including response quality, participant comfort, information richness, and technical reliability.
MimiTalk Research Team · 2025
Qualitative Researchers' Perspectives on AI Interview Platforms
An investigation into how qualitative researchers perceive and evaluate AI-powered interview platforms, examining opportunities and challenges in academic research contexts.
MimiTalk Research Team · 2025
AI Interviewers in Cross-Cultural Research Contexts
Exploring the effectiveness and adaptability of AI interviewers in cross-cultural research settings, examining cultural sensitivity and communication patterns.
MimiTalk Research Team · 2025
Applications of AI Interviewers in Educational Research
A conceptual paper examining the potential applications and implications of AI interview systems in educational research and assessment.
MimiTalk Research Team · 2025
Methodology guides
In-depth guides on qualitative research methods, written by our research team
Research from the field
Selected work on AI interviewing by other research teams — explore more on our blog