The central paradigm of Artificial Intelligence is rapidly shifting toward biological models for both robotic devices and systems performing such critical tasks as network management and process control. Here we apply recent mathematical analysis of the necessary conditions for consciousness in humans in an attempt to gain some understanding of the likely canonical failure modes inherent to a broad class of global workspace/blackboard machines designed to emulate biological functions. Similar problems are likely to confront other possible architectures, although their mathematical (...) description may be far less straightforward. (shrink)