A 5IR-inspired conceptual framework for AI-augmented construction estimation: review on developing conceptual BIM-integrated hybrid models for cost, time, and risk prediction

Abstract

Abstract This study introduces three hybrid AI conceptual models integrated with BIM for predictive analytics in construction management for future validation through simulations and ethical AI considerations by a systematic analysis of 66 studies. Despite the potential benefits of AI-BIM hybrid models for operational efficiency and accuracy, they remain underdeveloped and inadequately embody the principles of the fifth industrial revolution (5IR). The framework proposes three conceptual hybrid AI models for future validation: a dynamic cost predictor utilizing XGBoost and natural language processing for effective cost estimation; a schedule forecaster combining Long short-term memory networks with Monte Carlo simulations to enhance schedule predictions; and a risk analyzer that incorporates Bayesian networks and computer vision to evaluate multidimensional risks. These conceptual models are proposed to be interconnected with BIM to facilitate automated data extraction and provide real-time decision-making support. The proposed conceptual framework emphasizes ethical AI practices and promotes human-AI collaboration while aiming for sustainable resource management. A phased implementation roadmap was provided to guide pilot testing and industry adoption, highlighting the necessity for socio-technical integration. This proposed roadmap focuses on the seamless transition of these advanced methodologies into construction practices, ensuring robust collaboration among stakeholders. The proposed AI-BIM-5IR framework is conceptual and intended to guide future empirical validation and implementation studies. This research aims to present a foundation for future studies focused on enhancing resilience, accountability, and efficiency within the construction sector while adhering to the transformative principles of 5IR.