State-of-the-Art Review of Artificial Intelligence in Environmental Geophysics and Geotechnical Engineering

Artificial intelligence (AI) is transforming environmental geophysics and geotechnical engineering (EGGE), shifting practice from empirical and deterministic workflows toward data-rich, physics-consistent, and decision-oriented subsurface intelligence. This review synthesizes advances in machine learning, deep learning, physics-informed and theory-guided modeling, multimodal data fusion, uncertainty-aware and explainable AI, and intelligent sensing for near-surface, environmental, and geotechnical systems. It presents an integrated framework linking physics-informed AI, multimodal fusion, and regulatory pathways for deployment in EGGE, bridging methodological innovation and operational adoption. These advances enable high-resolution subsurface characterization, lithological and geotechnical profiling, hydro-geomechanical parameter estimation, groundwater assessment, geohazard forecasting, and infrastructure monitoring, marking a transition to field-validated decision-support systems for early warning, risk-informed design, and climate-resilient management. Key challenges include data heterogeneity, cross-scale inconsistency, limited ground truth, fusion complexity, ill-posed inversion, weak generalization across geological and climatic regimes, and insufficient interpretability, uncertainty quantification, and validation for engineering decisions. Regulatory gaps further constrain adoption. A roadmap emphasizes scalable physics-integrated AI, next-generation multimodal fusion, interpretable and uncertainty-aware modeling, edge-cloud digital twins, adaptive data acquisition, and standardized benchmarking. AI can redefine EGGE through resilient, sustainability-aligned subsurface intelligence if scientific rigor and governance advance in parallel.