<![CDATA[The convergence of Artificial Intelligence (AI) and Software Engineering (SE) has fundamentally reshaped how software systems are conceived, designed, built, tested, and maintained. Following the widespread commercial deployment of Large Language Models (LLMs) and generative AI tools such as GitHub Copilot, this paper presents a narrative review of current developments, key challenges, and future research directions in AI-driven Software Engineering (AI4SE). Drawing on peer-reviewed systematic literature reviews, empirical studies, and industry survey data published between 2019 and 2025, this review synthesizes evidence across the software development life cycle (SDLC), including requirements engineering, software design and architecture, automated code generation, software testing, and software maintenance. The review shows that AI adoption among practitioners has grown rapidly. Industry surveys report that more than 80% of developers now use AI tools in their workflow while empirical studies report productivity gains of up to 55% for specific coding tasks. However, the literature also converges on persistent challenges, including code quality and technical debt, explainability and trust, data and model bias, security vulnerabilities in AI-generated code, and organizational barriers to adoption. This review proposes an integrative conceptual framework that maps AI capabilities onto SDLC phases and challenge dimensions, and outlines an agenda for future research emphasizing trustworthy AI4SE, human-AI collaboration models, and empirical validation at scale. The paper is intended to serve as a reference synthesis for researchers, practitioners, and policymakers navigating the rapidly evolving intersection of AI and software engineering.]]>
