AI for Mass Timber Construction (MTC): Design, Optimization, Manufacturing, and Digital Twin Integration: A Scoping Review

Abstract

Abstract

Artificial intelligence (AI) is rapidly transforming engineering, manufacturing, and materials science, creating new opportunities to improve automation, sustainability, and lifecycle performance in the built environment. As demand for low-carbon construction increases, Mass Timber Construction (MTC) has emerged as a promising alternative to conventional structural systems due to its renewable nature, prefabrication potential, and reduced embodied carbon. However, despite growing interest in both AI and MTC, AI applications remain fragmented across multiple disciplines and lifecycle stages, and a comprehensive synthesis of current developments, challenges, and future research directions is lacking. In this scoping review, the authors systematically examine the current state of AI integration across the lifecycle of MTC systems, including material characterization, structural design, manufacturing automation, digital twins, structural health monitoring, and lifecycle sustainability assessment. Using the PRISMA Extension for Scoping Reviews (PRISMA-ScR) framework, relevant literature was identified, screened, and synthesized to evaluate emerging methodologies, technological trends, implementation barriers, and research gaps in AI-enabled MTC. The authors demonstrate that, unlike conventional construction materials, MTC often lacks the large and standardized datasets required for purely data-driven modeling, leading to increased adoption of scientific machine learning approaches, including physics-informed, mechanics-informed, and hybrid AI frameworks. Recent advances show that multimodal sensing techniques, including imaging, spectroscopy, acoustic sensing, and embedded monitoring systems, are increasingly being integrated with optimization algorithms and digital workflows to improve structural performance, vibration behavior, fire resistance, embodied carbon assessment, and manufacturing precision. In manufacturing, AI-enabled robotics and cyber-physical systems employing computer vision and reinforcement learning have enhanced lamella grading, machining accuracy, adhesive application, and process reliability in Cross-Laminated Timber (CLT) and Glued-Laminated Timber (glulam) production.  In building design, digital twin frameworks that combine physics-based simulation with real-time sensing enable the predictive assessment of moisture transport, deformation, vibration response, and long-term structural performance. The authors identify major challenges, including fragmented digital infrastructures, limited interoperability between design and simulation platforms, insufficient field-scale validation, fragmented multimodal data, and uncertainty in long-term model generalization. These limitations restrict the transfer of AI models across species, manufacturing environments, and operational conditions, while also hindering regulatory acceptance and large-scale implementation in safety-critical timber applications.  To address these challenges, the authors propose a unified lifecycle intelligence framework for AI-native MTC systems in which material behavior, manufacturing processes, and building-scale performance are connected through continuous data flows, hybrid physics-informed models, and cyber-physical feedback mechanisms. Building upon this framework, a future research roadmap is developed that prioritizes standardized multimodal datasets, interoperable digital infrastructures, scalable validation protocols, and explainable hybrid physics–AI methodologies to support the deployment of lifecycle-aware MTC systems.