1) There is no way to logically and factually prove BG3 is difficult to play. Everything required to play this game reasonably well is well within the reach of most game players who have acquired basic gaming skills.
Difficulty is a personal subjective experience. If someone finds the game too difficult to the point they stop playing it, they are right. If a significant portion of the player base feels and acts this way, then it is safe to say that the game is "factually" hard.
In chess, the AI is designed to win, to beat you. That's the goal.
In BG3, the AI isn't designed to beat you. It's designed for you to win. The only question is "how hard is it supposed to be for you to win?"
When I ask for a harder mode in the game, what I want is a game where the rules are the same for both sides and the AI is designed to beat you, to win. That way I'm going up against a real challenge.
Game developers never focus on making a super smart AI, because it is a massive amount of work for a questionable outcome. They usually have to cut out a large portion of the planned content anyway, due to time constraints, so the smart AI has no chance of getting anywhere near the backlog. There are cheaper ways to add difficulty: more enemies, that do more damage and last longer.
On the other hand, it is also very hard to balance the game around a smart AI. In the context of BG3, the result of a fight between an expert player and AI would boil down to initiative and pure RNG. The one who goes first, wipes the other one out, provided all attack rolls are successful. I don't see how this could be seen as fun. The randomness goes heavily against the skill aspect of the challenge. You need a fairly stupid AI to be able to build enough buffers to counteract the effects of a potentially catastrophic RNG event.
More realistically though, even a very smart AI would have a number of blindspots to be exploited, rendering the entire effort to create one pointless.
