Assessing the effectiveness of AI generated code in Improving Software Engineering Processes using GPT-Engineer with OpenAI GPT models

This paper evaluates GPT-Engineer, an open-source AI-powered tool leveraging OpenAI’s models, other Large Language Models (LLMs), and flexible integrations to enhance software engineering workflows. We applied GPT-Engineer to document an API codebase using Swagger Documentation, showcasing the tool’s potential to streamline engineering processes, optimize resources, and address challenges such as workforce shortages. This study focused on assessing GPT-Engineer’s effectiveness in generating precise, reliable API documentation and its understanding of the underlying codebase by targeting specific Laravel controllers and models, without requiring manual annotations. The evaluation explored adaptability, prompt effectiveness, and cost-efficiency using the "improve (-i)" module within GPT-Engineer across various GPT models. Findings indicate that AI-driven tools like GPT-Engineer can mitigate resource constraints, enabling organizations in low-resource environments to produce high-quality software. Key results highlight the importance of prompt engineering strategies, effective cost management, and scalable integrations for maximizing the utility of AI solutions in resource-limited settings.