Resilience-by-Design: AI for security, sustainability and health in interdependent systems

This title-driven scoping review synthesizes a conceptual framework for designing resilient AI-enabled systems across security, energy, industrial, health and social domains. Resilience encompasses anticipating disturbances, absorbing shocks, recovering quickly, and adapting to new realities. Modern systems are increasingly interdependent, and failures in one domain (e.g., cyber or energy) can cascade across sectors and threaten societal wellbeing. This review summarises the inferred sector, task, and methodological features from titles to map how AI capabilities support resilience functions. Methods such as forecasting, anomaly detection, predictive maintenance, optimization and explainable learning are linked to resilience functions and deployment settings (edge, cloud, 6G, federated). Titles also highlight trust dimensions—privacy, security, bias control and governance—that must accompany technical solutions. We develop a resilience-centric taxonomy, a cross-sector failure-mode synthesis, assurance checklists and reference design patterns suitable for practitioners. The synthesis emphasises that resilience is a socio-technical property: technical robustness must align with transparent governance, auditability, stakeholder collaboration and context-specific risk management. Because this review is based solely on titles, it provides a conceptual structure rather than evidence-based conclusions. The final sections outline a research agenda and limitations, and caution that full-text verification is necessary before drawing firm conclusions.