Can ChatGPT Generate Realistic Synthetic System Requirement Specifications? Results of a Case Study

TL;DR

This study explores ChatGPT's ability to generate realistic synthetic system requirement specifications (SyRSs) across industries, highlighting both its potential and limitations through systematic evaluation and expert feedback.

Contribution

The paper demonstrates a systematic approach for generating and assessing synthetic SyRSs with ChatGPT, revealing insights into their realism and the challenges involved.

Findings

62% of experts considered the SSyRSs to be realistic

Generated 300 SSyRSs across 10 industries using prompt patterns

LLM-based assessments cannot fully replace expert evaluations

Abstract

System requirement specifications (SyRSs) are central, natural-language (NL) artifacts. Access to real SyRS for research purposes is highly valuable but limited by proprietary restrictions or confidentiality concerns. Generating synthetic SyRSs (SSyRSs) can address this scarcity. Black-box large language models (LLMs) such as ChatGPT offer compelling generation capabilities by providing easy access to NL generation functions without requiring access to real data. However, LLMs suffer from hallucinations and overconfidence, which pose major challenges in their use. We designed an exploratory study to investigate whether, despite these challenges, we can generate realistic SSyRSs with ChatGPT without having access to real SyRSs. Using a systematic approach that leverages prompt patterns, LLM-based quality assessments, and iterative prompt refinements, we generated 300 SSyRSs across 10 industries with ChatGPT. The results were evaluated using cross-model checks and an expert study, with n=87 submitted surveys. 62\% of experts considered the SSyRSs to be realistic. However, in-depth examination revealed contradictory statements and deficiencies. Overall, we were able to generate realistic SSyRSs to a certain extent with ChatGPT, but LLM-based quality assessments cannot fully replace thorough expert evaluations. This paper presents the methodology and results of our study and discusses the key insights we obtained.