EXPLAINABLE ARTIFICIAL INTELLIGENCE AND EDGE COMPUTING FRAMEWORKS FOR SECURE DECISION MAKING IN SMART DIGITAL ECOSYSTEMS

Abstract

Artificial Intelligence is increasingly a key enabler of decision-making, with fast, adaptive, data-driven approaches to smart digital ecosystems in healthcare, smart cities, Industry 5.0, energy systems, IoT networks, and cyber-physical infrastructures. The increased and potential deployment of opaque AI models in distributed settings, however, comes with concerns about trust, security, accountability, privacy, and operational resilience. This review examines the convergence of explainable artificial intelligence and edge computing as a foundation for secure decision-making in smart digital ecosystems. It synthesizes the conceptual foundations of smart ecosystems, edge intelligence, trustworthy AI, and cybersecurity, followed by an analysis of major XAI techniques, including intrinsic models, SHAP, LIME, counterfactual explanations, attention-based methods, and knowledge graph-supported approaches. The review further discusses edge computing frameworks based on IoT–Edge–Cloud integration, federated learning, privacy-preserving computation, blockchain-enabled trust, and security-aware decision mechanisms. Real-world implementations are examined across critical application domains to show how explainable edge intelligence can improve transparency, responsiveness, and human confidence. The analysis identifies persistent challenges related to resource-efficient explainability, adversarial robustness, privacy preservation, model validation, governance, and context-aware deployment. In conclusion, this review underscores the importance of developing common standards and guidelines that incorporate explainability, security, edge intelligence, and human-centric governance to support clear and secure decision-making in the smart digital ecosystems of the future.