Accurately predicting risk or potentially risky situations is critical for the safe operation of autonomous systems. Risk refers to the expected likelihood of an undesirable outcome, such as a collision. We draw on an existing conceptualization of the risk to evaluate a robot’s options (e.g. choice of a path to travel). In this context, risk consists of two components: 1) the probability of an undesirable outcome computed by a Bayesian Network (BN) and 2) an estimate of the loss associated with the undesirable outcome. We demonstrate that our risk assessment tool is effective at computing the anticipated risk over a wide variety of the robot’s options and selecting the option with the lowest risk for two different types of autonomous systems: An Autonomous Vehicle (AV) operating near a college campus and a pair of Unmanned Aerial Vehicles (UAVs) flying from Washington DC to Baltimore. The method for assessing risk is used to identify higher risk routes, days to travel, and travel times for an autonomous vehicle and higher risk routes for a UAV.