Water distribution systems (WDS) are vulnerable to contaminants, and systematic risk assessment can provide valuable information for assisting threat management. Contamination events are sociotechnical systems, in which the interactions among consumers and water infrastructure may generate unpredicted public health consequences. This research develops a sociotechnical risk assessment framework that simulates the dynamics of a contamination event by coupling an agent-based modeling (ABM) framework with Monte Carlo simulation (MCS), genetic algorithm (GA) optimization, and a multi-objective GA. The ABM framework couples WDS simulation with agents to represent consumers in a virtual city. MCS is applied to estimate the uncertainty in human exposure, based on probabilistic models of event attributes. A GA approach is used to identify critical contamination events by maximizing risk, and a multi-objective approach explores the trade-off between consequence and occurrence probabilities. Results that are obtained using the sociotechnical approach are compared with results obtained using a conventional engineering model. The sociotechnical approach removes assumptions that have been used in engineering analysis about the static, homogeneous, and stationary behaviors of consumers, and results demonstrate new insight about the impacts of these actions and interactions on the public health consequences of contamination events.