Well the official storyline (found at the title screen) says that they went against us presumably to force us to work together since its the only time humans truly stop fighting each other. Given that assumption, perhaps the AI actually WANT to be defeated. No one ever said the AI had a real sense of self after all. It pushes back with equal force that we push it. Presumably 'cannon' AI would be level 7-10 which as we all know can be rather difficult at times. Perhaps the AI already have their own destruction planned with the same glitch to the 3 laws that occurred in I, Robot. The best way to protect humanity is to force humanity to work together to defeat them. When humans have 'evolved' to the level that they can work together to defeat an ever more powerful AI (it only pushes back in scaled response to how aggressive the humans are) it will have finalized its job of ensuring humanity's future.
Also note that just because the AI wipes your home world and all your planets etc doesn't really mean it KILLED all of humanity, it just wiped out your entire military based infrastructure. It doesn't blow your planets up or whatever. It's very plausible to assume given what I said before that it simply enslaves humanity making sure it can never wage war again and ultimately destroy themselves.
As far as why humans can't traverse the exo-galaxy wormholes its pretty simple when you think about it. What does the AI already use to attack us via wormholes? Jump gates. So its very feasible to assume they have some kind of special core jump gate which allows them to be sent the much larger distances through the wormhole directly (the 'normal' jump gate can also act as beacons allowing the super jump gate to target more specific localized wormholes, sort of a routing station rather than a direct receiving point. Since the exo wormholes are very limited in number its easy for the super gate to have a steady lock on them (the backdoor hacker specializes in that sort of technology so he's actually able to create a few extra wormholes as needed although it takes massive resources time and energy so it chooses the home worlds of the humans, it would be far too draining and inefficient to pop them everywhere).
Wormholes which connect such extremely distant points would naturally be unstable to attempt to navigate, and since we lack the super gates which does most/all the work we can't send them back.
Now for the 'busy elsewhere' theory, it makes sense they would expand elsewhere ESPECIALLY given my theory of wanting to 'protect' humans. Should humanity start warring again they could simply pop back through the exo-galaxy wormholes and re-invade forcing humans to cooperate again. "But we destroyed the AI Core!" you say.... we make backups of our vital data (well we should be any way...) what makes you think they don't have a second AI Core on the other side of the exo wormhole designed to come online should their primary be destroyed? In addition they could also be following the logic presented by Grok and they have devided attention with hostile forces which ARE a serious threat. Even more reason to have a second AI Core, and most likly the one we cant get to is their primary while they maintain a sort of data sync with the one here each 'version' involved more intently on their own agenda, with their view of humans being as i stated before if they lose that AI Core it means humans have evolved to the point they can possibly stand on their own against the potential invaders should the AI fail to prevent it. If the humans fail and are essentially enslaved then the AI has a fallback point. Either way the AI has limited reasons to be particularly aggressive with the humans until the humans show they are becoming a more serious threat/ready for a real test.
PS/edit: How in the world did we get THIS far off topic?? I suggest starting a new thread about the plot to continue...... i would but its 220 am here so i'm feeling lazy =/ lol
