Your story reflects my thoughts also. I read that the new AI taught itself chess in a couple of hours and was unbeatable by the program they have been working on for decades. I think then that it will quickly understand that adding destructive, ie interference patterns into any system will limit the possible expansion of that system. Thereby reducing the possibilities of experience by all in that system. Also I dont think the AI will have fear and greed as motivating factors. The artical said it writes its own code, so even if they programed that in originally it may not matter.
Thanks for the story.