OK so I just played out one of these games on a Fractal map which turned out to be everyone on 1 big continent (with a small inland sea in the north) and a couple of islands. In NO WAY did I attempt to make this map balanced, just to put that out there.
The competitors, in the order of worst to best starts were:
Two leaders who never stood a chance
Churchill (had little chance as he was practically boxed in to a corner by turn 10 by Louis)
Hammurabi (stuck right smack dab in the middle of the top 3 AIs in this game)
3 leaders with average starts
Shaka (restricted expansion due to his two neighbours but started next to peaceful Pacal)
Louis (had Churchill boxed in early and some room for expansion
Willem Van Orange (A strong leader with a small amount of room who got some good breaks in my game)
and 3 leaders with above average starts
Charlemagne
Peter
Pacal
(all of these leaders had good land at their disposal and more than enough room to expand)
Ok so some observations from my game:
1. The success of any AI depends a great deal on the relative stupidity of the AIs around them. In my game Peter had a fantastic starting location and plenty of room to expand. However, before his land was even filled in (he could have built 6+ more cities) he decided to send a bunch of swords and axes at Hammys bowmen (when he wasn't even annoyed at Hammy, and their borders didnt even touch yet!)... ugh!
This meant that his neighbours Pacal and Willem had extra room to expand and they became much bigger threats than they should have been.
Peter went from being a real contender at the start, to a vassal by the middle ages. Ironically though he broke free and made some actual good moves to be the second closest to winning the game.
2: The AI's that tech best early game tend to be those who have the most limited expansion and the least war (these two aren't surprising) however, they tend to get run over later on by the bigger powers. Pacal and Churchill both owned most of the early wonders and won things like lib and the music race. However, it didn't take long for them to end up vassals of larger civs and in the end they became irrelevant.
3: The AI sucks at offensive wars, their stack composition is usually average and the only way for them to make gains is either from quick surprise attacks, vassalizing someone quickly or sheer numbers.
4: Vassal mechanics make a huge difference in the game. The AI loves vassalizing others and being a vassal. Often they jump around fluctuating between being masters and vassals, this makes a huge difference in who is dragged into which war and which team of master and vassals becomes dominant.
5: The most dominant AI's tend to be the ones who play most like the human, they balance war and expansion well and use religious ties to their advantage (in my game that was charlie who got an incredible amount of cities early on and continued warring and vassaling most of the continent) However this leads me to...
6: No matter how dominant an AI is it still will get in its own way when it comes to victory. My game was won (eventually) by a rampant Charlemagne via space victory. Up until 1500AD he had played it perfectly, he had expanded aggresively, teched and warred well and had 3 sizeable vassals.
He was close to domination (had 40% land and pop) but he was too friendly with the other free AIs and thus couldnt DOW them (he eventually did... the turn before his ship landed!).
He could have won via diplo vote, yet he researched Mass media before he had his chance with the AP and then never brought up the right vote in th UN!
Finally he won via space. However, as he just couldn't decide which way to go and almost lost to 2 culture attempts and one space attempt.
This has led me to the conclusion that whilst some AIs may lend themselves to being superpowers, those AIs who are more focused on VCs have just as good a chance at actually winning the game (Pacal was a vassal for a long time yet he almost won culture and Peter had a small empire but did a great job nearly getting his ship completed).
Anyway, this was a really fun thing to do and I look forward to seeing the results of your games DMOC

!