He is not satisfied, and maybe because of this it will happen.
perfect extermination of the enemy - yes, for this is (Golem) has been produced, but what did happen?
I don't see the destroying potential in this look into the future, only his intellect, as million as fast as ours.
Sogo, it looks like you're the great fan of AI. I may be wrong, but this is what goes straight from your sentences quoted above.
OK, Golems predecessors were indeed produced for one purpose - to eliminate the enemy. It was the time of the Cold War. They were, if I remember well, fully featured AI machines. But they had some internal blockades limitating their free will. Golem XIV was the first one with erased blockades, so he (or it) was able to develope his own self-consciousness, emotions, feelings and so on...
He became maybe the greates philosopher we ever had. Finally it turned out that his thoughts were too difficult for us, more over it was not that easy for Golem himself to share his thoughts with us. He found our human language too limited for knowledge sharing process.
OK, now I think that Lem's look-aheads are not related to his emotions, as you stated it in the first sentence of your answer. Lem is in some part a scientist, philosopher and who knows what else...
And this is why his visions of future comes true. His fiction is based on the very science.
As I mentioned before, I haven't read Summa yet, but I read Megabit Bomb (that's my translation of the book's title (in polish - Bomba Megabitowa), that was not yet translated into english - Terminus - thanks for your remind) and I figured out that Mr. Lem's attitude to the latest techno-scientific discoveries is negative.
So if you believe you don't see a destroying potential in AI it's because we still don't know how to build it. At the moment it's just beyond of our reach. Experiments are in progress but there are no results so far.
CU,
Deckard