Answers are grounded in this paper and cite their sources.
Trust remains insufficiently understood in construction AI, particularly where AI outputs inform safety-critical or contractual decisions with professional and legal consequences. This study investigates how construction professionals evaluate and justify trust under technical uncertainty, site variability and professional accountability. A cross-national mixed-methods design was used to analyse 1,233 valid survey responses from professionals in managerial, technical, design and site-based roles across Australia, the United Kingdom and the United States. The study distinguishes trust calibration, defined as aligning trust with an AI system’s capabilities and limitations, from calibrated professional reliance, whereby AI outputs are relied upon only when they are technically reliable, contextually appropriate and professionally defensible. Across the three countries, accuracy, reliability and contextual alignment formed the basis for reliance, while governance, explainability, transparency and accountability supported reliance under safety, contractual and regulatory scrutiny. The main risks concerned unreliable predictions and data privacy and security, while the main benefits were efficiency, planning accuracy and decision support. The study advances trust-in-AI research by conceptualizing calibrated professional reliance as a construction-specific judgement shaped by technical performance, contextual fit and professional accountability. These findings explain how construction professionals decide whether to accept, verify, modify, override or reject AI outputs.
No takes yet. Share an insight, caveat, or question.
Regona et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: