Artificial Intelligence with a precise research focus that signals what makes this review distinct from existing meta-analyse

Abstract

Artificial Intelligence (AI) has emerged as one of the most transformative technologies of the 21st century, permeating diverse domains including healthcare, education, business, manufacturing, and environmental science. This study presents a rigorous mixed-method review integrating PRISMA 2020 guidelines with bibliometric mapping using VOSviewer and Bibliometrix (R package). A total of 1,247 articles were initially retrieved from Scopus. After a four-stage PRISMA screening process, 183 articles were retained for in-depth analysis. Bibliometric analysis revealed that the United States, China, and the United Kingdom are the leading contributors. The most productive journals include Nature Machine Intelligence, Artificial Intelligence Review, and Expert Systems with Applications. Co-citation analysis identified five major thematic clusters. Keyword co-occurrence analysis highlighted "deep learning," "machine learning," "neural networks," "NLP," and "explainable AI" as the most prominent terms. The findings reveal significant research gaps including limited attention to AI fairness, underrepresentation of developing-country perspectives, and insufficient longitudinal studies.