Risk analysis of implementing Machine Learning in construction projects

Sammanfattning: Machine Learning has significantly influenced development across domains by leveraging incoming and existing data. However, despite its advancements, criticism persists regarding its failure to adequately address real-world problems, with the construction domain being an example. Construction sector is crucial for global economic growth, yet it remains largely unexplored, lacking sufficient research and technological utilization of its extensive data. Despite increasing publications on adapting technological advancements, the primary focus is on urging industry innovation through digitization. Recently, adopting Machine Learning to address operational challenges has gained attention. While some studies have explored potential ML integration opportunities in construction, there is a gap in understanding the factors and barriers hindering its adoption across projects. This study investigates the factors restricting organizations from implementing ML in construction projects and their consequent operational impacts. This study employs a comprehensive literature review of ML concepts and identifies gaps in construction data. Qualitative interviews have been conducted in a semi-structured manner with five industry professionals offering practical insights, preceding a thematic analysis of interview data. Themes are analyzed and discussed in relation to theoretical material to identify connections. Finally, a risk assessment based on identified risks is evaluated through a risk matrix. The results of this study discuss the challenges and potential benefits of implementing ML within the construction industry. The study further emphasizes the necessity of knowledge to understand project-specific datasets. With a primary focus on unstructured text and image data, the study uncovers challenges related to data inconsistency that affect data reliability. While recognizing ML’s potential to streamline construction operations, the study underscores challenges such as data security and digitalization. In summary, this study emphasizes the importance of data quality, security, and cultural transformation in harnessing ML’s capabilities to improve construction project management and operations.