Document Type : Review paper
Authors
1
Southern Delta University, Ozoro
2
Ministry of Justice, England, United Kingdom
3
Department of Mechanical Engineering, Faculty of Engineering, Southern Delta University, Ozoro, Delta State, Nigeria
4
Agricultural Engineering, faculty of engineering, Southern Delta University, ozoro, Delta state, Nigeria
10.22059/jchpe.2026.405358.1684
Abstract
ABSTRACT
Artificial Intelligence (AI) is changing the concept of process safety management by facilitating predictive and prescriptive control instead of reactive power. The old safety systems consist of manual checks and lagging indicators that cannot reflect the nonlinear behavior of complex industrial processes. This systematic review summarizes the results of 91 peer-reviewed publications (2020-2025) that report on the use of AI and Machine Learning (ML) for predictive hazard detection and decision support in the chemical, manufacturing, transportation, and energy industries. There are five major areas of application: AI-assisted hazards prediction, machine-learning-powered fault detection, real-time monitoring via deep learning, digital-twin-powered predictive maintenance, and hybrid physics-AI modeling. The use of AI has shown decisive advantages in the timely identification of anomalies, predictive maintenance, and risk-based decision making, and is the basis of intelligent Safety 4.0 frameworks. Nonetheless, obstacles still exist, such as a lack of data, model interpretability, cybersecurity risks, and the difficulty of integrating AI and legacy infrastructure. The way forward lies in developing explainable AI (XAI), hybrid modeling, and human-AI cooperation to enhance transparency, trust, and ethical governance. The review finds AI to be a fundamental technology in predictive, resilient, and human-centered process safety in industries being transformed toward Industry 5.0.
Keywords
Subjects
