Seismic activity can trigger extreme events with severe impacts on the environment, populations, and infrastructure, particularly in communities with limited response capacity. Disaster risk management is crucial and should be integrated into broader development strategies. Disaster Risk Reduction (DRR) frameworks help identify hazards and vulnerabilities to enhance preparedness. Over the past decade, efforts to assess seismic vulnerability at an urban scale have advanced, particularly in historic centers of seismically active regions such as Italy, Chile, and Portugal, where index-based methodologies have improved risk management for built heritage. However, these methodologies are highly sensitive to uncertainty, which can lead to inaccuracies in risk assessments. To address these challenges, this work proposes a novel machine learning classification algorithm that leverages artificial intelligence to streamline the survey and data collection process for well-known vulnerability parameters specific to unreinforced masonry, typically found in historical structures. These parameters include number of floors, aggregate position and interaction, plan configuration, wall façade openings and alignment, and non-structural elements. Automating data acquisition reduces both the time required for data collection and the uncertainty in seismic vulnerability assessments of historical buildings. This methodology was applied to over 700 historical buildings of the Chimba quarter in Santiago (Chile), with the results analyzed using an integrated Geographical Information System (GIS). The integration of AI and seismic vulnerability assessment enhances disaster preparedness and risk mitigation, contributing to a more resilient and sustainable future for vulnerable historical regions.