# Can AI Serve as a Substitute for Human Subjects in Software Engineering Research?

**Authors:** Marco A. Gerosa, Bianca Trinkenreich, Igor Steinmacher, Anita Sarma  
**arXiv:** [2311.11081](https://arxiv.org/abs/2311.11081)  
**Date:** 2023-11-18  
**Summary page:** https://www.summarizepaper.com/en/arxiv-id/2311.11081v1/

## Key points

- Sociotechnical domains like Software Engineering face challenges with qualitative data collection methods in terms of scale, labor intensity, and participant recruitment.
- The proposed solution is to leverage artificial intelligence (AI), specifically large language models (LLMs) such as ChatGPT, for qualitative data collection in software engineering research.
- AI-generated synthetic text can replicate human responses and behaviors, enabling automation of data collection across various methodologies like persona-based prompting for interviews, multi-persona dialogue for focus groups, and mega-persona responses for surveys.
- AI models offer scalable and efficient means of data generation while providing insights into human attitudes, experiences, and performance.

## Summary

, , , , 


In the realm of sociotechnical domains like Software Engineering, qualitative data collection methods face challenges in terms of scale, labor intensity, and participant recruitment. To address these issues, this vision paper proposes leveraging artificial intelligence (AI), specifically large language models (LLMs) such as ChatGPT, for qualitative data collection in software engineering research. By utilizing AI-generated synthetic text that replicates human responses and behaviors, researchers can automate data collection across various methodologies like persona-based prompting for interviews, multi-persona dialogue for focus groups, and mega-persona responses for surveys. The paper discusses how AI models could offer scalable and efficient means of data generation while providing insights into human attitudes, experiences, and performance. : In the realm of sociotechnical domains like Software Engineering

: Qualitative data collection methods face challenges in terms of scale, labor intensity, and participant recruitment. : This vision paper proposes leveraging artificial intelligence (AI), specifically large language models (LLMs) such as ChatGPT. : By utilizing AI-generated synthetic text that replicates human responses and behaviors. : For qualitative data collection in software engineering research.