Evolution emerges from random mistakes, not intentional design.

Evolution up to humans has, as far as we are able to observe, evolved from apparently random variations. The word mistake implies intention, which requires intelligence.

Of course, it is possible that intelligence and goal seeking has been baked into the universe from the beginning. But I don't think you were implying that.

Anyway, now that we can all agree that the universe does have intelligence, in at least the form of humans, it's obvious that further intelligence evolves from a combination of pre-existing intelligence and random variations, which we can now logically call mistakes, since there is definitely intention at play in creating the environment for these mistakes to come into being.

And I don't think an AI will just emerge from the blockchain without further intelligent intervention by humans. Which is of course happening.

Besides all that, what will be the goal of this AI? And would we even be able to determine that it exists and what its goals are?

I mean, consider the idea that the earth or even the cosmos are intelligent systems. We have no way of testing these hypotheses. We don't know what goals they could possibly be seeking. We don't know their language. Hell, we have a hard enough time just trying to communicate with dolphins and whales, which we know have language.

So, I find it within the realm of possibility that an AI could emerge without us ever knowing it. If it were a truly intelligent and autonomous being, completely separate from us, having none of the same needs, such as food, oxygen, companionship, etc., would it even have the need to communicate with us? Will its thoughts be in a language completely indecipherable and seemingly nonexistent to us?