Abstract
Artificial intelligence (AI) has increasingly transformed architectural and urban research through data-driven modelling, spatial analytics, and decision-support systems. Despite this growth, limited attention has been given to human-centred perspectives, particularly children’s spatial experiences, well-being, and ergonomic needs in public environments. This imbalance raises questions about whose needs are represented within AI-supported urban research. This study examines the intersection of AI, architecture, urban design, and public space through bibliometric analysis and a PRISMA-guided systematic review. An initial dataset of 871 publications from Web of Science and Scopus was screened, resulting in 39 studies. A supplementary child-oriented search identified 11 studies, of which four addressed child-related spatial considerations. Findings reveal that existing research is dominated by urban analytics, machine learning, and performance-oriented approaches, while child-centred perspectives remain underrepresented. Comparison between AI-assisted clustering and manual thematic classification shows that automated approaches capture dominant trends but often overlook socially relevant marginal themes. Beyond this gap, the study proposes a preliminary Child-Centred AI Urban Design Framework (CCAI-UDF) linking AI-driven spatial analysis with human-centred architecture and child-responsive design. The framework advances a transition from efficiency-driven AI toward inclusive child-centred urbanism, contributing to future AI-supported ergonomic assessment tools, urban policy, and equitable public space design.