"If intelligent machines were to take over the world because they were much better at performing human processes, would this count as evolution?"
Answer 1. It's all a semantic issue that is almost entirely arbitrary.
Answer 2. In the most general definition of the term 'evolution', yes, it is evolution.
Answer 3. Is it biological evolution? I'm not sure.
An sf book you might find interesting: "Code of the Lifemaker," by Hogan.
In any case, I don't expect anything like this (computers that are "intelligent" in the ordinary sense that we mean the term) to happen in my lifetime and probably not in my children's lifetimes. Possibly in my grandchildren's lifes, but I doubt even that. One day, I think we'll figure it out, but not too soon.
Recent news:
Stanford won DARPA's grand challenge this year. They did a great job and they probably deserve the $2 million. But I'm guessing those conditions were relatively controlled. It's significant progress, but it's still only a fraction of where we'd like to be with autonomous vehicles (or so I'm guessing).
A japanese company recently demonstrated a robot that could ride a bike and stop and start without falling over. Significant, but we're still not there. I think we need numerous quantum leaps in technology to get there from here.
