Teoría y práctica del desarrollo de software aumentado con IA en Europa en 2025
TL;DR
The findings confirm the research hypothesis that AI can reliably produce key software development artifacts with minimal human input, marking a pivotal step toward the broader integration of AI in software engineering practices.
Abstract
This article explores the practical application of AI tool (ChatGPT v3.5 and 4) in the full lifecycle of a real-world software development project, based on field research of 2023-2024 in Europe. The research evaluates the effectiveness of AI in generating core software artifacts, including functional specifications, software code, automated tests, user documentation, and application content. While the AI tool proved to be a powerful assistant for creating and refining these deliverables, its limitations became evident in areas requiring up-to-date technical guidance or creative input in product business logic development. In general, ChatGPT significantly accelerated the development process, with project team members reporting a substantial increase in productivity. The study also highlights best practices for AI usage in software engineering, emphasizing the importance of service-oriented design, iterative prompt refinement, and ongoing human oversight. Despite its shortcomings in generating original ideas or adapting to evolving platform requirements, ChatGPT demonstrated strong capabilities in automating repetitive tasks and enhancing overall efficiency. The findings confirm the research hypothesis that AI can reliably produce key software development artifacts with minimal human input, marking a pivotal step toward the broader integration of AI in software engineering practices.