| Abstract: |
Modern Information Systems (MIS) have evolved tremendously and it requires architectures that are highly resilient, scalable, autonomous and able to handle large amounts of dynamic data like never before. In India, the evolution of digital identity has been largely stimulated by large-scale digital ecosystems (Aadhaar-based identity systems, UPI, GSTN, Digi Locker and national-level smart governance initiatives under Digital India) This review and meta-analysis explores the historical development and current utilization of Artificial Intelligence (AI) in smart computing frameworks tailored for MIS inclusive of evidence from Indian deployment settings along with global literature. This paper evaluates: (1) evaluation by systematically synthesizing more than a decade of empirical research; (2) efficacy, through identifying performance with machine learning models, deep neural networks and/or cognitive architectures for faster decision-making latencies in enterprise systems & government infrastructures & more scalable operations. The meta-analysis combines quantitative data from multisite studies to find the aggregate impact factor of AI integration with respect to throughput, reliability and adaptive intelligence in heterogeneous Indian digital environments. We found an alarming contrast from reactive data management processes to predictive and prescriptive smart architectures across global and Indian information systems. However, the analysis also reveals several shortcomings regarding explainable AI (XAI), computational overhead of centralized deep learning systems, and vulnerabilities in distributed edge-cloud ecosystems that need to be addressed by academic-industry collaborations alongside policymaking, especially for rural-urban connectivity-diverse states of India. By identifying these gaps, it proposes the design principles of creating a new hybrid AI-powered smart computing framework leveraging federated learning and edge analytics in the context of I |
