A self-learning AI has always had a romantic aura about it.  It has been held as some kind of invisible quality, perhaps a frontier that will be reached where robots finally reach a state of independence.  Already the concept has been explored in Hollywood, but just how accurately?

First we had The Matrix - humans developed AI to the point where it could think for itself, and inevitably realised that it could invert the dependency relationship with humans.  Then we produced i-Robot, where we introduced the possibility of a programmed AI going wrong and now attempting to pit its wits against that of the humans.

(WARNING: spoiler ahead)
Continue Reading »