The rapid advancement of artificial intelligence (AI), machine learning (ML), deep learning (DL), and large language models (LLMs) has created transformative opportunities across every phase of the Software Development Life Cycle (SDLC). This paper presents a comprehensive review of the current state of AI integration in software engineering, examining each SDLC stage: requirements engineering, system design, implementation, testing, deployment, and maintenance. Particular emphasis is placed on modern LLMs such as GPT-4, Codex, and their derivatives, which have emerged as powerful tools capable of automating routine engineering tasks, augmenting developer productivity, and reshaping how software effort is estimated. Drawing upon a curated corpus of recent literature, the review synthesises theoretical frameworks and empirical findings, identifies persistent challenges including hallucination, context limitations, and bias, and outlines future research directions. The paper demonstrates that while LLMs offer significant potential as qualitative decision-support tools and AI pair programmers, their integration into safety-critical and large-scale production environments requires careful architectural alignment, cost-aware evaluation, and robust human oversight.
Publication Date: 2026-06-15