AI-DRIVEN PREDICTIVE HSE AND ENVIRONMENTAL RISK MANAGEMENT FRAMEWORK FOR SUSTAINABLE OIL AND GAS AND CONSTRUCTION PROJECTS IN SAUDI ARABIA

Abstract

Saudi Arabias oil and gas and construction sectors are expanding and digitalising, while facing increased expectations for occupational safety, process safety, environmental protection, and sustainable project delivery. This review introduces an integrated, AI-driven predictive health, safety, and environment (HSE) and environmental risk management framework designed for these high-risk environments. Using a structured integrative approach, the review synthesises recent peer-reviewed evidence on machine learning, deep learning, computer vision, IoT sensing, digital twins, predictive maintenance, natural-language processing, and decision-support systems. The evidence shows that AI can shift HSE management from periodic inspections and lagging indicators to continuous risk sensing, probabilistic forecasting, dynamic prioritisation, and earlier intervention. In construction, research is most advanced in visual hazard recognition, PPE monitoring, injury prediction, and digital safety workflows. In oil and gas, research has advanced asset integrity, pipeline risk assessment, methane monitoring, corrosion prediction, and incident analysis. However, technical accuracy alone does not guarantee safer or more sustainable outcomes. Ongoing challenges include data quality, class imbalance, interoperability, model drift, explainability, privacy, cybersecurity, governance, and human oversight. This review proposes a five-layer framework that integrates multisource data acquisition, data governance, predictive analytics, risk intelligence, and controlled intervention, with continuous feedback to improve models and organisational learning.