Because many interconnected networks are locally self-organizing, each is built with concern for local constraints and optimizations, and each is on a separate growth schedule, the problem of securing and robustifying these interlocked systems cannot itself help but deal with the global or architectural features of these systems as they are themselves networked. Networked networks have recently been shown to carry additional risks of cascading collapse; for example, scale-free networks are individually very robust, maintaining their integrity even when confronted with numerous random errors, but sets of similarly structured networks interconnected through their nodes can be far less stable and can experience catastrophic multinetwork cascades as a result of far fewer errors. These findings conflict with the common intuition that multinetworked systems would be more stable because of the potential for built-in redundancy. The implications for maintaining security in nested systems are twofold. Failures in individual systems can rapidly lead to global multisystem failures, and, consequently, precautions for preventing a catastrophe in one system will require monitoring and regulating other connected systems — potentially on a global scale. This session on precaution in networked networks will feature lectures reflecting on the current state-of-the-art in this area of network robustness.