Trustworthy AI in Sustainable Building Projects: Prioritizing Data Quality for Risk Management Decisions

Artificial intelligence (AI) is increasingly being adopted for decision support in sustainable building risk management, yet the trustworthiness of AI-supported sustainability risk decisions depends as much on data quality as on analytical capability. Poor data conditions can amplify sustainability risks by producing unreliable decision support, yet existing studies provide limited insights into which data quality dimensions should be prioritized to enable trustworthy AI outcomes. This study identifies and prioritizes the critical data quality dimensions for trustworthy AI-supported decisions in sustainable building risk management. A questionnaire survey was conducted of accredited sustainable building professionals and their expert judgements were analyzed through an Analytic Hierarchy Process (AHP). The findings reveal that system-dependent dimensions, particularly traceability and interoperability, are prioritized over intrinsic dimensions like accuracy and consistency. The findings suggest that trustworthy AI-supported sustainability decisions depend strongly on a verifiable data provenance, cross-system integration and interpretable outputs rather than data correctness alone. This study reframes data quality from a general prerequisite to a prioritized, context-sensitive construct underpinning trustworthy AI applications, extending data-driven decision theory in the sustainable building domain. Ultimately, a phased data governance approach is recommended to prioritize traceability and interoperability as the foundational conditions for construction organizations implementing trustworthy AI in sustainable building risk management.