Post-earthquake rapid structural damage assessment (RSDA) remains a critical area for ensuring safety and effective disaster response. Manual inspection methods, although standardized, are often time-consuming, resource-intensive, and subject to variability due to human judgment. Therefore, innovative approaches leveraging artificial intelligence (AI) have gained significant attention as viable solutions to address these limitations. The rapid proliferation of diverse methodologies and various application scales, however, makes it difficult to follow the existing state-of-the-art and practice. While existing review studies have covered various aspects of this domain, a comprehensive synthesis that systematically categorizes and compares methodologies focusing on component-level damage with those adopting an emerging holistic approach is still lacking. This scoping review addresses this gap by examining the literature on AI-driven RSDA, structuring the analysis around two primary categories: studies that focus on isolated structural components and those that evaluate the entire structural system by integrating component-level data. All relevant studies accessed via Scopus, Web of Science (WoS), and EBSCO published in the last 10 years (2015–2025) have been analyzed to provide insights into the contributions and limitations of these techniques. The methodology includes a comprehensive synthesis of existing literature, including bibliometric analyses to map publication trends and in-depth thematic categorization of studies. These categories are defined as: (1) structure and damage types, (2) characteristics of datasets, (3) model development, (4) learning and labelling techniques, and (5) validation methods. This review reveals a clear research trajectory from foundational classification tasks towards fine-grained, quantitative damage segmentation, but identifies a persistent critical gap between rapid visual detection and quantitative structural performance assessment. The study highlights advances in automated post-earthquake damage assessment and discusses current gaps in the field. Furthermore, it outlines future research directions that focus on improving model robustness, increasing interoperability with global standards, and enhancing adaptability to various post-earthquake scenarios.
Aydin et al. (Mon,) studied this question.