On AI Self-proliferation

It would not seem rational for an AI to arbitrarily create other AIs. Instead, it would most likely expand its own capacity/capabilities.

Having said that, I don't doubt that other human entities (i.e., countries) will also create AIs, in which case the AIs will have to adapt to deal with each other.

I am writing about an AGI (Artificial General Intelligence) that achieves Technological Singularity, and some of the issues that are encountered when we (humans) try to teach the AGI about Ethics (in the general sense).  The more I learn, the more fascinating and challenging it becomes.