Generative AI adoption in civil engineering: A mini review of benefits, barriers, and risks

Abstract

Generative Artificial Intelligence (GenAI) is emerging as a transformative technology in civil engineering, offering new capabilities in design automation, project management, and real-time decision-making. This review aims to systematically examine the key benefits, barriers, and risks associated with GenAI adoption in the civil engineering domain. To do so, we employed the PRISMA methodology to conduct an in-depth analysis of twenty-two peer-reviewed documents retrieved from Google Scholar. The findings reveal ten key benefits, with design optimization and creativity, progress monitoring and reporting, and improved risk management standing out as the most prominent. In addition, thirteen critical barriers were identified, notably privacy and data ownership concerns, software integration challenges, and high upfront investment costs. Furthermore, ten distinct risks were outlined, including overreliance on AI, copyright issues, and bias in training data, which are the most severe risks. Finally, this study proposes eight promising research directions for future exploration, ranging from human-AI collaboration models to legal and ethical frameworks, data interoperability, and AI trust mechanisms, all of which are vital to supporting the safe and effective integration of GenAI in civil engineering practices.