I am no computer expert at all but for what i understood :
Before AI software tried to use trees, recognition of shapes. … but that didn’t work well because too many possibilities.
With AI of today they introduce the use of probability, and a way to learn to play better by learning by itself. The last point is essential and the main research, it makes computer play better as human because it will be more efficient in his way of processing. They even proved that they don’t need human historic knowledge of the game, computer will discover all and more by itself.
There are articles to read, they should better explain as me! (Search deepmind/alphago )
Note that all this is not so useful to elaborate a strategy to use to win against a computer by anticipating the moves he will play, it’s just not possible.
The programmer foresee nothing. Even the software don’t really approach the game move by move but generate tons of fullboard position with the basic rules and then apply probability on these. (Called Monte Carlo)
Stronger players may envy you because they have much less freedom as you have when chosing the move to play !
Happy gaming.