In my opinion, artificial intelligence is not very suitable for poker. One of the main things in poker is the ability to bluff. How well and convincingly can an artificial intelligence bluff? It may well learn to do so over time, though.
A bluff is deliberate false information.
Here it is worth asking a much broader question.
And actually, in percentage terms, AI can provide false information to a person. And how will AI evaluate how much false and how much objective information to provide to a person in communication with him? In my opinion, AI now has so much false information that it itself is no longer able to distinguish the truth from an obvious lie. Well, if this is so, then the AI will probably learn such a poker tool as bluffing very quickly. And most importantly, it is completely unclear how to try to recognize this bluff.
This is not a rival, a person who still has some emotions that can be read on his face or by his gestures or something else visually.
And this will just strengthen the AI’s position as an opponent in the game. He becomes a more dangerous opponent.