The escalating frequency and intensity of seismic events, compounded by rapid urbanization, pose a significant threat to global transport infrastructure. Traditional seismic hazard assessment and structural engineering methods, while valuable, often struggle with the non-linear, high-dimensional, and complex nature of earthquake phenomena and soil-structure interactions. The advent of Artificial Intelligence (AI) and Machine Learning (ML) heralds a paradigm shift, enabling a transition from reactive response to predictive engineering. This paper provides a comprehensive review of the integration of AI and ML methodologies—including Remote Sensing, GIS, Information Value, Frequency Ratio, Logistic Regression, Artificial Neural Networks, and advanced deep learning architectures—into seismology for the safeguarding of transport infrastructure. We synthesize how these technologies are revolutionizing seismic hazard prediction, ground motion characterization, liquefaction susceptibility mapping, and real-time structural health monitoring. The review critically analyzes the capabilities of various ML models, presents their applications through summarized case studies, and discusses the challenges of model interpretability, data scarcity, and integration into engineering practice. Finally, we outline future research directions, emphasizing the potential of physics-informed neural networks and digital twins to create a robust, predictive, and resilient framework for transport infrastructure in seismically active regions.