From data to policy: a systematic review of AI in water regulations and compliance

Abstract

Artificial intelligence (AI) algorithms are increasingly recognized for their transformative potential in water regulation and management. This systematic survey reviews 100+ peer-reviewed studies that applied AI-driven techniques to advance water policy across three critical policy domains: water pollution, drinking water, and infrastructure regulation. Using a novel multi-dimensional evaluation framework, this systematic review assesses the methodological rigor, policy integration, real-world applicability, AI assurance, and the socioeconomic or ethical impact of each reviewed study. The results reveal that AI applications in water governance are primarily dominated by machine-learning and neural-network methods, focusing on pollution control and infrastructure optimization. In contrast, far fewer studies employ causal, reinforcement-learning, or decision-support approaches that directly engage with policy evaluation. Despite technical advancements, only 18% incorporated assurance elements such as explainability, fairness, or trustworthiness, underscoring the need for more transparent and deployable AI in water governance. This review synthesizes global research efforts and outlines key opportunities and challenges. In response, it outlines future research directions that emphasize policy-aware decision-support systems, the advancement of trustworthy and explainable AI assurance frameworks, and the systematic incorporation of socioeconomic and ethical considerations to strengthen accountability and equity in water governance.