While yes you are right, that some games will only see some graphical improvement, AI will become vastly improved in the next generation.

In FPS games, an AI would be more intelligent, making human-like decisions and reacting in a human-like manner as situations change.  Rather than the more predictable routines AIs use today.  It would represent a fundamental change in the play of the game.  In RPGs, an AI would be capable of running real-time scenarios for non-player characters making the choice of when and how quickly you take on a quest vitally important.  Unlike the way most RPGs function today in forcing quests to be played one after the other or queue up and completed at your own speed.  In sports games, an AI would be capable of portraying the unique playing styles of each player, while also possessing the ability for the NPCs on both teams to react in an even more life-like manner to their position and the ball.

This level of AI sophistication wouldn't be possible on either the 360 or the PS3 as the CPUs are already overburdened.  The graphical improvements of the next generation are only a small bit of what gamers will see.  More realistic AI will likely be the one they notice first and foremost.   Though, that likely will take a year before we see games truly taking advantage of the power available to programmers. 

If anything, third parties will go the same way THQ did.

ahahahaha

Ok the PS4 will be as powerful as the PS3.

Are we seeing said AIs on PC games, then, if CPU is the limiting factor? I think the limiting factor is budget and dev time, which will actually be more strained by the requirements of next-gen. I don't think better AI in an FPS would be a reason to buy a game unless it was transformatively better.

Sports games have huge incentive to develop a better AI already because it's a struggle to differentiate from last year's iteration, and the fact that none of them have tells me it must be harder than simply having a few more CPU cycles free.

It's a difficult decision for 3rd parties to just ignore a 140m install base. MS and Sony won't let 3rd parties to just ditch their new consoles in favor to the PS3/360, lots of money will be wasted for that not to happen, although given development cost, I expect this next gen transition to be longer than previous gens (even with the moneyhat).

That might happen in future generations, but i see the next generation being the same as previous generations. People will just adopt the new consoles gradually and buy games for it (1st or 3rd party), just like before

If you take a sample of uninformed high school gamers and put two TVs side by side with different levels of the same game running on each, PS3 and PS4, I don't think they'd be able to tell which one is the next-gen. But they certainly would for PS -> PS2 -> PS3. I don't mean they won't have more execution resources - they will - it's that current games already allow the full experience the developer intended to be shown, and extra resources will just go to more AA, more polygons, higher resolution, and other things you actually have to be looking for to recognise.

What I want to get across is that it will be slow, and painful. Next-gen won't be 'saving' anything like some people are implying. The current bad sales aren't waiting for next gen they're just bad.

If the reasons you mentioned came true, I'm pretty sure Microsoft/Sony can pull the plug on them.

so does 1st party will kill current gen at the same time... so end up no more PlayStation and Xbox in future... Nintendo F.T.W...