IMO the only thing "less info" will do is make it less obvious how bad AI is.
Fortunately this is a testable hypothesis. Load up obfusCKate, pick someone unfamiliar that you can't assume can just bulldoze anyone in the first years, and see if you play Crusader Kings 3 the same way.
If you play the same way, then perhaps all it does it makes it less obvious how bad the AI is. We can have a separate discussion about whether this is itself a Good Thing. Various aspects of the AI that people claim to like from CK2 were equally incompetent AI. However, it was often obfuscated behind complex mechanics that didn't make it obvious. AI army management is an example- the AI would never aim to optimize commanders and tactics like a player could. However, players who didn't understand how the tactics worked didn't mind. The system wasn't 'better' because the AI was actually good at it, just less-obviously bad.
If obfusCKate
changes how you play the game the game, however, then this premise is false. The way it changes how you play the game, in turn, can drive other AI balance decisions.
For example- a common player complaint is that it's too easy to go from count to emperor in a single generation, and then they get bored. A significant part of that comes from knowing which fights they can pick, and which they should delay, to start building their titles.
If the player does not know that by default- if the player has to spend
time and resources figuring out who is a safe target- then the time to conquer up takes longer. Which, in turn, makes it no longer a necessarily single-generational assumption.
Which, in turn, makes inter-generational mechanics- such as succession management and new ruler penalties- matter more than they would for a game which is functionally over before the first succession occurs. Instead of focusing balance on 'what keeps the first player character from rising,' balance can factor in things like 'what makes the player spend more time on managing the realm rather than wars to expand it.'
Or we could balance issues that could set-back the player in their first generation against different information limits. For example- more secret independence factions, who don't offer huge lead-up indicators, would be *far* more disruptive if they could revolt at lower thresholds (not 80%) and win quicker (say immediate ticking war score) if the player's MAA are some ways away in a war of aggression. This sort of thing wouldn't work on perfect-information mode when you know revolts are coming and how to defang them without putting in any work.