Interacting with cyclists is a difficult challenge for Automated Vehicles (AVs). It is important that the introduction of AVs does not lead to a decrease in cyclist safety. This study presents a new methodology that is developed to assess road safety between AVs and cyclists. Surrogate Measures of Safety are identified specifically for the interaction between AVs and cyclists. The methodology makes use of the presented Surrogate Measures of Safety to objectify road safety, by classifying each into three different safety levels. The methodology is presented in a rubric that can be used to assess road safety of conflict situations. The methodology is demonstrated for a case-study in the Netherlands, analyzing 163.5 km of road with a total of 118 conflict situations. For each conflict situation, road safety is assessed and accident risk factors are identified. Common accident risk factors found are reduced visibility of a cyclists’ trajectory path, unclear infrastructure due to missing road surface markings, and infrastructure facilitating visual communication and occasional cyclist traffic violations. Although safety primarily needs to be on-board of AVs, road authorities can improve road safety by providing clear and predictable infrastructure, with good visibility of conflict situations and cyclists’ trajectory paths. Future research should focus on improving the performance of cyclist detection, classification, and trajectory prediction by AVs. Potential conflict situations between AVs and cyclists should be analyzed in more detail to quantify their safety level, and AVs should be trained in conflict situations where cyclists can be encountered.