At present many researchers are exploring ways to improve the resilience of the global system, prevent catastrophic failure, and support the transition to a sustainable global system. This paper makes two interlinked proposals for developing better-and more credible-tools for modeling critical global issues and identifying and managing threats. First, proven risk management methods from other fields can be usefully applied to the assessment and mitigation of large, complex socio-ecological problems. Second, using viability as the priority criterion for the design of our models will not only highlight systemic threats, but also help us develop constructive interventions.