Abstract: Automated Vehicles (AVs) have strong potential to improve road safety by reducing human error, the leading cause of crashes. Yet, the distribution of these benefits across geographic and sociodemographic groups remains unclear. This study develops a framework to assess disparities in AV-related safety gains using Alabama and Kentucky as case studies. The approach combines: (1) modeling AV adoption interest with survey data and binary logit estimation, (2) incorporating spillover effects through inverse distance weighting (IDW), and (3) conducting a spatial crash analysis normalized by roadway length. Results show higher adoption interest among younger and higher-income groups, while rural residents express lower interest. Crash patterns vary spatially: urban areas experience more rear-end and sideswipe crashes, whereas severe rollover crashes are more common in rural regions. Comparing adoption interest with crash exposure reveals that rural communities, with higher severe crash rates, show lower adoption interest, underscoring the need for targeted policies.