<![CDATA[Purpose of the study: This review critically evaluates AIoT-based digital monitoring architectures that integrate edge intelligence and Water Quality Index (WQI) forecasting, in order to identify how distributed media-technology infrastructures can support real-time, predictive, and policy-aligned water quality monitoring across urban and rural environments. Methodology: A PRISMA-guided critical review was combined with VOSviewer 1.6.20 bibliometric mapping. Peer-reviewed articles indexed in Scopus, Web of Science, IEEE Xplore, ScienceDirect, MDPI, and SpringerLink between 2015 and 2025 were systematically screened, quality-appraised using a six-criterion rubric, and thematically synthesised, complemented by a quantitative nine-dimensional technical-performance comparison framework across cloud, edge, hybrid, and federated architectures. Main Findings: Three persistent weaknesses were identified: urban-centric architectural bias, supervised-learning dependence incompatible with rural data scarcity, and weak alignment between AI analytics and regulatory indices. Bibliometric clustering revealed four dominant research themes which is IoT sensing, machine-learning forecasting, edge intelligence, and federated/adaptive analytics. A hybrid edge-cloud AIoT framework with quantitative performance benchmarks is proposed to resolve these gaps. Novelty/Originality of this study: Unlike prior reviews that treat smart water monitoring as a uniform technical problem, this study reframes it as a distributed media-technology challenge, introduces a bibliometric-supported urban-rural taxonomy of AIoT architectures, and explicitly maps AI model placement onto the six-parameter Malaysian WQI computation, converting predictive analytics into a regulatory decision-support instrument with documented system orchestration and technical workflow.]]>
