The role of Artificial Intelligence in improving construction site safety

Abstract

Construction sites consistently rank among the most hazardous work environments worldwide, contributing a disproportionate share of occupational injuries and fatalities. This paper provides a comprehensive review of how artificial intelligence (AI) technologies are being leveraged to enhance safety in construction, synthesizing findings from engineering, computer science, safety management, and related fields. The discussion spans current AI applications – including computer vision for real-time hazard detection, robotics and autonomous systems for risk elimination, machine learning-based predictive analytics for accident prevention, and natural language processing for safety knowledge mining – as well as future directions and global implementation challenges. Real-world examples from diverse regions illustrate AI’s potential, such as computer vision systems monitoring worker compliance and predictive algorithms that forecast incidents before they occur. Quantitative findings from recent studies are highlighted, including reductions in accident rates and improved detection accuracy, with statistical tables summarizing key data. While AI offers powerful new tools for mitigating construction risks (e.g., identifying unsafe conditions, augmenting worker decision-making, and removing personnel from harm’s way), this review also addresses hurdles such as data quality, worker acceptance, regulatory concerns, and the need fo,r integrated, “human-in-the-loop” safety management approaches. The paper concludes that AI, when responsibly implemented, represents a transformative adjunct to traditional safety practices – capable of significantly improving construction site safety performance globally – but it must be deployed in tandem with organizational commitment, worker training, and robust safety cultures.