ARCHIVES

Generative Artificial Intelligence (GenAI), driven by Large Language Models (LLMs) and multi-modal deep learning architectures, has evolved from a novelty assistant to a core operational layer across the Software Development Lifecycle (SDLC). This paper provides a comprehensive analysis of the real-world applications, underlying mechanisms, and operational risks associated with integrating Generative AI into modern software engineering practices. We propose a six-stage lifecycle framework—ranging from requirements engineering to maintenance—and map specific Generative AI mechanisms (e.g., retrieval-augmented generation, fine-tuned transformer models, synthetic dataset generation) onto each phase. Furthermore, we synthesize empirical benchmark data to evaluate productivity gains against software quality, security, and architectural drift. The paper concludes with actionable mitigation strategies for legal compliance, security vulnerability remediation, and hallucination management in production-grade repositories.