I can say, back on the matter of top civs, that cultural one are really weak in these versions. I think in these versions, Ai can not pursue cultural victory, at least on king dif. because i play my games on communitas map, the world is very crowded, even with minimum watter lvl set and some civs just are in better shape to deal with that than the others. But from my point of view, civs, which does not get any relevant bonus to science( no matter source) are usually bottom. That's why Ethiopia, Maya, Assyria(they are about top3-4 civs in my games now), Ottoman and some others are usually a top AI civs. Some of them scyrocket higher with game progressing, some fall behind based on circumstances. But in my last game Maya, who picked Sainthood and was able to spread into 4 other civs, he was almost unstopable.
Maya seem to me best Ai controlled civ right now. Most bothering and very aggresively pursue whatever path it chooses. Settle aggresively, spread aggresively, not best at fighting but save because of military advancement, etc. This civ has some advantages and AI knows, how to use them. On the other hand, there are many civs, which have never reached their top and it seems to me, that they even does not know how.
For example japan. Very weak AI civ in my last 3 previous games. They always seem to pursue for some wonders and building advancement before they start expansion. With such crowded map, i think they could change their priority for their own good. Also bonus from dojo building, while very interresting for player and absolutely no initiative by me to change it, seem very weak for AI. I think AI could get yields from that trippled for that beign at least near efficient like for player. Before that change, where they did not have yields from military/defensive buildings, they have played just fine in a matter of early expansion. Now i saw them 3x in a row, first build a walls in a capital and probably barracks too. By that time other AIs have already settled 3-4 cities.