Abstract:

Machine learning (ML) has the potential to support the analysis of complex soil–structure interactions, improve predictive accuracy and enhance decision-making in geotechnical engineering. However, the integration of ML into geotechnical engineering raises critical concerns regarding trustworthiness, explainability, and accountability, particularly given the high-stakes nature of infrastructure design and construction. This presentation examines the role of ML in geotechnical practice through the lens of these three interrelated dimensions. The importance of uncertainty quantification, model interpretability and ethical implications of ML applications in geotechnical engineering are addressed, and the State-of-Practice is exemplified with several real-life case studies. The necessity of combining domain knowledge with responsible ML practices to ensure reliable, transparent and ethically sound adoption of ML in geotechnical engineering is also highlighted and discussed.