Artificial intelligence in process safety: A review of opportunities, challenges, and future directions for the chemical process industries

The Chemical Process Industries (CPIs) are undergoing a data-driven transformation, as artificial intelligence (AI), the Industrial Internet of Things (IIoT), and advanced analytics are redefining critical safety operations. This review consolidates recent applications of AI for hazard identification, dynamic risk assessment, early incident detection, and proactive barrier management in refineries and petrochemical facilities, emphasizing reported enhancements of 30–60 % in anomaly detection and near-miss classification compared to traditional methodologies. Ongoing challenges include sparse and biased incident datasets, limited interpretability, difficulties in validating algorithms under actual plant conditions, and regulatory ambiguities that limit widespread deployment. To mitigate these deficiencies, the paper proposes a multilayered framework that aligns AI functions with process safety management (PSM) elements, including asset integrity, alarm management, and emergency response. Future research priorities include the development of explainable, certifiable artificial intelligence, digital-twin-driven predictive risk assessment, and the integration of functional safety and risk governance standards, such as International Electrotechnical Commission (IEC) 61511 and ISO 31000. The review delineates pathways to establish safer, cleaner, and more resilient chemical and energy systems globally.