Exploring the Role of ChatGPT in User Research: Opportunities and Limitations

Introduction

The rapid popularity of ChatGPT has prompted many industries to explore how it can reduce costs and improve efficiency, and user research is no exception. This article provides an initial exploration of ChatGPT’s role in user research, concluding that while it cannot fully replace researchers, it shows strong auxiliary potential.

ChatGPT Overview

ChatGPT, an OpenAI conversational model, achieved over 100 million monthly active users within two months, outpacing platforms like TikTok and Instagram. Major tech companies (Microsoft, Google, Baidu, Alibaba) have integrated or are developing similar models. Its strengths lie in language understanding and generation, though it has limitations such as weak mathematical ability and verbose responses.

Can ChatGPT Replace User Researchers?

ChatGPT cannot fully replace user researchers, but it can assist in four key areas: (1) efficient content integration as a search engine, (2) a versatile “assistant” that drafts research frameworks and summarises interview notes, (3) a polishing tool for improving written content, and (4) a simulation tool for conducting or role‑playing interviews.

1. Efficient Content Integration (Search Engine)

Unlike traditional search engines that return multiple results, ChatGPT can synthesize information into concise answers, helping researchers quickly grasp concepts such as the definition, technical features, and value of digital collectibles (NFTs). However, users must verify the factual accuracy of its responses, as the model can produce confident but incorrect statements and cannot cite specific sources.

ChatGPT also offers guidance on data‑analysis methods (e.g., Excel formulas, TGI index) but shows limited support for specialized tools like SPSS.

2. Versatile “Assistant” (Content Extraction)

ChatGPT can extract key information from interview transcripts and questionnaire data, providing initial research frameworks, interview outlines, and summaries. For example, it successfully identified core insights from three user statements about digital collectibles, such as the importance of platform credibility and content differences across platforms.

It can also compare non‑real statistical tables and suggest design recommendations, though deeper insight still requires human interpretation.

3. Content Polishing Tool

ChatGPT can correct typographical errors and improve English phrasing. The article shows examples where it fixes Chinese misspellings and rewrites a Chinese sentence into more elegant English, demonstrating usefulness for polishing interview guides, questionnaire items, and reports.

4. Interview Simulator

By prompting ChatGPT to act as an interview host or as a typical user, researchers can generate interview scripts or simulate user responses. The model asks structured questions about NFTs and can adopt personas (e.g., a 20‑year‑old college student) to provide plausible answers, offering a way to prototype interview flows before real user sessions.

Conclusion

ChatGPT can serve as a cost‑effective assistant across multiple user‑research stages: rapid knowledge retrieval, summarising qualitative data, polishing written material, and simulating interview interactions. However, its outputs still require expert validation, fine‑tuning for the research domain, and better interaction designs before it becomes a fully reliable tool.

Tips for Practitioners

1) Experiment multiple times to compare answers; 2) Craft clear, incremental prompts; 3) Maintain critical thinking and verify the truthfulness of AI‑generated content.