Theoretical and practical instantiations of Generative AI in construction risk management: an analytical exposition of its latent benefits and inherent risks

The construction industry's increasing complexity and dynamic project environments engender advanced risk management strategies. AI-based risk management tools, reliant on complex mathematical models, often impose specialised coding requirements, leading to challenges in accessibility and implementation. In this vein, Generative Artificial Intelligence (GenAI) emerges as a potentially transformative solution, leveraging adaptive algorithms capable of real-time data analysis to enhance predictive accuracy and decisionmaking efficacy within Construction Risk Management (CRM). However, integrating GenAI into CRM introduces significant challenges, including concerns around data security, privacy, regulatory compliance, and a skills gap. Our research seeks to address these issues by presenting a systematic bibliometric analysis that explores evolving trends, key research contributions, and critical methodological approaches related to GenAI in CRM. Thus far, our investigation has analysed 23 selected research articles from an initial corpus of 212 papers, spanning the period from 2014 to 2024. Early insights delineate a marked escalation in research activity from 2020 onwards, a surge likely engendered by recent advancements in AI technologies and their applicability to construction management. We categorise GenAI's potential benefits into technical, operational, technological, and integration-related advantages, encompassing improvements in risk identification, predictive capabilities, scheduling, and cybersecurity. Simultaneously, we identify significant risks, particularly related to data governance, social acceptance, and the operational impacts of AI-driven decisions. These preliminary findings underscore the imperative for systematic governance frameworks and proactive stakeholder engagement to optimise GenAI's benefits whilst mitigating its latent risks.