Can ChatGPT exceed humans in construction project risk management?

Abstract

Purpose: The objective of this research is to investigate the capabilities of the ChatGPT GPT-4 model, a form of artificial intelligence (AI), in comparison to human experts in the context of construction project risk management. Design/methodology/approach: Employing a mixed-methods approach, the study draws a qualitative and quantitative comparison between 16 human risk management experts from Finnish construction companies and the ChatGPT AI model utilizing anonymous peer reviews. It focuses primarily on the areas of risk identification, analysis, and control. Findings: ChatGPT has demonstrated a superior ability to generate comprehensive risk management plans, with its quantitative scores significantly surpassing the human average. Nonetheless, the AI model's strategies are found to lack practicality and specificity, areas where human expertise excels. Originality/value: This study marks a significant advancement in construction project risk management research by conducting a pioneering blind-review study that assesses the capabilities of the advanced AI model, GPT-4, against those of human experts. Emphasizing the evolution from earlier GPT models, this research not only underscores the innovative application of ChatGPT-4 but also the critical role of anonymized peer evaluations in enhancing the objectivity of findings. It illuminates the synergistic potential of AI and human expertise, advocating for a collaborative model where AI serves as an augmentative tool, thereby optimizing human performance in identifying and managing risks.