I don't expect much. I have yet to see any AI for any strategy game as open-ended as Civ that wasn't hopelessly exploitable and predictable. I'm sure Civ VI will be the same way.
But the designers can help themselves out by making design decisions that are AI-friendly. Civ V really shot itself in the foot with the new 1UPT system that put a ton of stress on the AI's tactical combat abilities. The AI was of course hopeless at handling its armies (what else could the designers have expected?) and that had deleterious effects on the whole game. If the AI could never really threaten you because it handled its armies so incompetently, then the game is less tense, diplomacy is less important, conquest is less satisfying, etc.
A further problem was that there was no really easy way to give bonuses to mask the AI's total incompetence. Yes, because of production bonuses the AI typically had more troops, but troops are useless if you don't know what to do with them. And the designers chose not to take drastic measures that would have made combat more difficult, but much less fair--for example, having AI troops get big combat bonuses against the player, or heal more, or move faster, etc. So the AI just gets steamrolled in combat.
On the other hand, for the empire-building side of the game, the bad AI is less of a big deal. If Civ has a really rich and interesting economic system that the AI can't make heads or tails of (as in Civ IV), that's no big deal--just give it enormous production and research boosts (that are invisible to the player and don't feel so unfair), and it will be perfectly competitive.
I doubt very much that Civ VI will have a decent combat AI. And, just as in Civ V, that might be the main thing that keeps it from becoming a truly great title.
