Generative Artificial Intelligence in Software Engineering Applications, Challenges and Future Perspectives

Abstract

Generative Artificial Intelligence (GAI) has emerged
as a transformative technology in software engineering, redefining established practices in code generation, automated
testing, documentation, and quality review. Tools such as GitHub
Copilot, ChatGPT, and Large Language Models (LLMs) have
demonstrated the capacity to accelerate development cycles,
reduce routine errors, and democratize advanced technical skills.
However, their widespread adoption raises critical questions
about the security of generated software, ethical responsibility,
data privacy, and technological dependency. This paper presents
a systematic analysis of GAI applications in software engineering,
quantifies their advantages and limitations through published
empirical evidence, and examines ethical and security considerations relevant to their deployment in production environments.
Results indicate that GAI can reduce coding time by up to 55.8%,
but introduces security vulnerabilities in 40%–60% of generated
code without human supervision. It is concluded that responsible
adoption requires governance frameworks, automated vali