Socio-technical determinants of explainable artificial intelligence for infrastructure decision support.

Area of Science:

• Infrastructure Management
• Artificial Intelligence
• Explainable AI (XAI)

Background:

• Growing AI use in infrastructure improves decision-making but faces trust issues due to algorithmic opacity.
• Limited research exists on socio-technical factors for trustworthy explainable AI in infrastructure.
• Existing studies focus on predictive performance, neglecting trust-building conditions.

Purpose of the Study:

• Examine socio-technical determinants of trust in explainable AI.
• Investigate factors influencing perceived infrastructure decision quality.
• Understand how explainable AI impacts infrastructure management decisions.

Main Methods:

• Quantitative research design.
• Structured questionnaire survey with 283 infrastructure professionals.
• Partial Least Squares Structural Equation Modeling (PLS-SEM) analysis.

Main Results:

• Algorithmic transparency, perceived explainability, and decision accountability significantly enhance decision quality.
• Transparency, explainability, and accountability strengthen trust in explainable AI.
• Organizational digital readiness did not significantly impact trust in explainable AI.

Conclusions:

• Effective AI-supported infrastructure decision-making hinges on aligning explainability with organizational governance.
• Trust in explainable AI is built through transparency, explainability, and accountability.
• Socio-technical factors are crucial for successful explainable AI adoption in infrastructure management.