This paper reports progress towards the development of a practical risk analysis and management framework for intelligent information systems based on the state-of-art techniques in uncertainty management. We provide an analysis of challenges raised by the need to manage risk and identify a set of key requirements for a practical framework that can support risk management in real environments that are open, complex and dynamic. We assess a number of relevant theories, approaches and techniques for their suitability in addressing the risk management challenges. Finally, we present our current multi-level risk analysis and modelling framework, and use benchmark problems in two entirely different domains to illustrate the broad range of our framework applicability.