Explainable Artificial Intelligence: Precepts, Methods, and Opportunities for Research in Construction

Construction industry lags behind other sectors in adopting explainable AI: The paper notes that while XAI has gained traction in various industries, construction has given it "limited attention," suggesting the sector risks falling behind in responsible AI implementation compared to other fields.

Explainability can reduce skepticism about AI in construction: The authors argue that XAI opportunities could "alleviate the scepticism and hesitancy toward AI adoption and integration in construction," indicating that making AI decisions transparent and understandable is key to building trust among construction stakeholders.

Different stakeholders have different needs for AI explanations: The paper identifies "stakeholder desiderata" as a critical research opportunity, suggesting that construction workers, managers, clients, and other parties may require different types of explanations to understand and accept AI systems in their work.

Better data integration requires clearer AI reasoning: The paper highlights "data and information fusion" as an important research area, implying that as construction AI systems become more complex by combining multiple data sources, explaining how these systems reach conclusions becomes increasingly important.

A structured framework for XAI in construction is needed: By developing "a taxonomy of the XAI literature comprising its precepts and approaches," the authors suggest the construction industry needs organized guidance on XAI methods rather than adopting AI explanations haphazardly.